update:替换编译器自带的全部内存管理，格式化打印及部分时间函数，消除使用自带的malloc等内存管理函数可以带来的安全隐患

add:看门狗中断时的提示
fix:debug模式下没有终止看门狗复位

Third_Party/ZBAR/src/qrdectxt.c

@@ -314,7 +314,7 @@ int qr_code_data_list_extract_text(const qr_code_data_list *_qrlist,
             cur_eci=entry->payload.eci;
             if(cur_eci<=QR_ECI_ISO8859_16&&cur_eci!=14){
               if(cur_eci!=QR_ECI_GLI0&&cur_eci!=QR_ECI_CP437){
-                sprintf_(buf,"ISO8859-%i",QR_MAXI(cur_eci,3)-2);
+                sprintf(buf,"ISO8859-%i",QR_MAXI(cur_eci,3)-2);
                 enc=buf;
               }
               /*Note that CP437 requires an iconv compiled with

Third_Party/cm_backtrace/cm_backtrace.c

@@ -163,7 +163,8 @@ static size_t main_stack_size = 0;
 static uint32_t code_start_addr = 0;
 static size_t code_size = 0;
 static bool init_ok = false;
-static char call_stack_info[CMB_CALL_STACK_MAX_DEPTH * (8 + 1)] = { 0 };
+//static char call_stack_info[CMB_CALL_STACK_MAX_DEPTH * (8 + 1)] = { 0 };
+extern const uint8_t ByteToAsciiTable[16];
 static bool on_fault = false;
 static bool stack_is_overflow = false;
 static struct cmb_hard_fault_regs regs;
@@ -401,29 +402,23 @@ size_t cm_backtrace_call_stack(uint32_t *buffer, size_t size, uint32_t sp) {
  * @param sp stack pointer
  */
 static void print_call_stack(uint32_t sp) {
-    size_t i, cur_depth = 0;
+    size_t i,j, cur_depth = 0;
     uint32_t call_stack_buf[CMB_CALL_STACK_MAX_DEPTH] = {0};
-
     cur_depth = cm_backtrace_call_stack(call_stack_buf, CMB_CALL_STACK_MAX_DEPTH, sp);
-
+    char call_stack_info[CMB_CALL_STACK_MAX_DEPTH * (8 + 1)] = { 0 };
     for (i = 0; i < cur_depth; i++) {
-        #ifdef __BUILD_APP__
-        sprintf_(call_stack_info + i * (8 + 1), "%x", call_stack_buf[i]);
-        #else
-        sprintf(call_stack_info + i * (8 + 1), "%08lx", call_stack_buf[i]);
-        #endif
+    	for (j = 0; j < 8; j++)
+    	{
+    		call_stack_info[i * (8 + 1) + j] = ByteToAsciiTable[(unsigned char)((call_stack_buf[i] >> (28-j*4))&0x0f)];
+    	}
         call_stack_info[i * (8 + 1) + 8] = ' ';
     }
-
     if (cur_depth) {
         cmb_println(print_info[PRINT_CALL_STACK_INFO], fw_name, CMB_ELF_FILE_EXTENSION_NAME, cur_depth * (8 + 1),
                 call_stack_info);
     } else {
         cmb_println(print_info[PRINT_CALL_STACK_ERR]);
     }
-#ifdef __BUILD_OS__
-    Core_PrintServiceStack();
-#endif
 }
 
 /**

Third_Party/cm_backtrace/cmb_def.h

@@ -297,7 +297,7 @@ struct cmb_hard_fault_regs{
 if (!(EXPR))                                                                   \
 {                                                                              \
     cmb_println("(%s) has assert failed at %s.", #EXPR, __FUNCTION__);         \
-    while (1);                                                                 \
+    return;                                                               \
 }
 
 /* ELF(Executable and Linking Format) file extension name for each compiler */

Third_Party/cm_backtrace/en-US/cmb_en_US.h

@@ -0,0 +1,71 @@
+/*
+ * This file is part of the CmBacktrace Library.
+ *
+ * Copyright (c) 2020, Armink, <armink.ztl@gmail.com>
+ *                     Chenxuan, <chenxuan.zhao@icloud.com>
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining
+ * a copy of this software and associated documentation files (the
+ * 'Software'), to deal in the Software without restriction, including
+ * without limitation the rights to use, copy, modify, merge, publish,
+ * distribute, sublicense, and/or sell copies of the Software, and to
+ * permit persons to whom the Software is furnished to do so, subject to
+ * the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be
+ * included in all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED 'AS IS', WITHOUT WARRANTY OF ANY KIND,
+ * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.
+ * IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY
+ * CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT,
+ * TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE
+ * SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * NOTE: DO NOT include this file on the header file.
+ * Encoding: UTF-8
+ * Created on: 2020-09-06
+ */
+
+[PRINT_MAIN_STACK_CFG_ERROR]  = "ERROR: Unable to get the main stack information, please check the configuration of the main stack",
+[PRINT_FIRMWARE_INFO]         = "Firmware name: %s, hardware version: %s, software version: %s",
+[PRINT_ASSERT_ON_THREAD]      = "Assert on thread %s",
+[PRINT_ASSERT_ON_HANDLER]     = "Assert on interrupt or bare metal(no OS) environment",
+[PRINT_THREAD_STACK_INFO]     = "===== Thread stack information =====",
+[PRINT_MAIN_STACK_INFO]       = "====== Main stack information ======",
+[PRINT_THREAD_STACK_OVERFLOW] = "Error: Thread stack(%08x) was overflow",
+[PRINT_MAIN_STACK_OVERFLOW]   = "Error: Main stack(%08x) was overflow",
+[PRINT_CALL_STACK_INFO]       = "Show more call stack info by run: addr2line -e %s%s -a -f %.*s",
+[PRINT_CALL_STACK_ERR]        = "Dump call stack has an error",
+[PRINT_FAULT_ON_THREAD]       = "Fault on thread %s",
+[PRINT_FAULT_ON_HANDLER]      = "Fault on interrupt or bare metal(no OS) environment",
+[PRINT_REGS_TITLE]            = "=================== Registers information ====================",
+[PRINT_HFSR_VECTBL]           = "Hard fault is caused by failed vector fetch",
+[PRINT_MFSR_IACCVIOL]         = "Memory management fault is caused by instruction access violation",
+[PRINT_MFSR_DACCVIOL]         = "Memory management fault is caused by data access violation",
+[PRINT_MFSR_MUNSTKERR]        = "Memory management fault is caused by unstacking error",
+[PRINT_MFSR_MSTKERR]          = "Memory management fault is caused by stacking error",
+[PRINT_MFSR_MLSPERR]          = "Memory management fault is caused by floating-point lazy state preservation",
+[PRINT_BFSR_IBUSERR]          = "Bus fault is caused by instruction access violation",
+[PRINT_BFSR_PRECISERR]        = "Bus fault is caused by precise data access violation",
+[PRINT_BFSR_IMPREISERR]       = "Bus fault is caused by imprecise data access violation",
+[PRINT_BFSR_UNSTKERR]         = "Bus fault is caused by unstacking error",
+[PRINT_BFSR_STKERR]           = "Bus fault is caused by stacking error",
+[PRINT_BFSR_LSPERR]           = "Bus fault is caused by floating-point lazy state preservation",
+[PRINT_UFSR_UNDEFINSTR]       = "Usage fault is caused by attempts to execute an undefined instruction",
+[PRINT_UFSR_INVSTATE]         = "Usage fault is caused by attempts to switch to an invalid state (e.g., ARM)",
+[PRINT_UFSR_INVPC]            = "Usage fault is caused by attempts to do an exception with a bad value in the EXC_RETURN number",
+[PRINT_UFSR_NOCP]             = "Usage fault is caused by attempts to execute a coprocessor instruction",
+#if (CMB_CPU_PLATFORM_TYPE == CMB_CPU_ARM_CORTEX_M33)
+    [PRINT_UFSR_STKOF]        = "Usage fault is caused by indicates that a stack overflow (hardware check) has taken place",
+#endif
+[PRINT_UFSR_UNALIGNED]        = "Usage fault is caused by indicates that an unaligned access fault has taken place",
+[PRINT_UFSR_DIVBYZERO0]       = "Usage fault is caused by Indicates a divide by zero has taken place (can be set only if DIV_0_TRP is set)",
+[PRINT_DFSR_HALTED]           = "Debug fault is caused by halt requested in NVIC",
+[PRINT_DFSR_BKPT]             = "Debug fault is caused by BKPT instruction executed",
+[PRINT_DFSR_DWTTRAP]          = "Debug fault is caused by DWT match occurred",
+[PRINT_DFSR_VCATCH]           = "Debug fault is caused by Vector fetch occurred",
+[PRINT_DFSR_EXTERNAL]         = "Debug fault is caused by EDBGRQ signal asserted",
+[PRINT_MMAR]                  = "The memory management fault occurred address is %08x",
+[PRINT_BFAR]                  = "The bus fault occurred address is %08x",

Third_Party/cm_backtrace/zh-CN/cmb_zh_CN.h

@@ -0,0 +1,71 @@
+/*
+ * This file is part of the CmBacktrace Library.
+ *
+ * Copyright (c) 2020, Armink, <armink.ztl@gmail.com>
+ *                     Chenxuan, <chenxuan.zhao@icloud.com>
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining
+ * a copy of this software and associated documentation files (the
+ * 'Software'), to deal in the Software without restriction, including
+ * without limitation the rights to use, copy, modify, merge, publish,
+ * distribute, sublicense, and/or sell copies of the Software, and to
+ * permit persons to whom the Software is furnished to do so, subject to
+ * the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be
+ * included in all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED 'AS IS', WITHOUT WARRANTY OF ANY KIND,
+ * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.
+ * IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY
+ * CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT,
+ * TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE
+ * SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * NOTE: DO NOT include this file on the header file.
+ * Encoding: GB18030
+ * Created on: 2020-09-06
+ */
+
+[PRINT_MAIN_STACK_CFG_ERROR]  = "错误：无法获取主栈信息，请检查主栈的相关配置",
+[PRINT_FIRMWARE_INFO]         = "固件名称：%s，硬件版本号：%s，软件版本号：%s",
+[PRINT_ASSERT_ON_THREAD]      = "在线程(%s)中发生断言",
+[PRINT_ASSERT_ON_HANDLER]     = "在中断或裸机环境下发生断言",
+[PRINT_THREAD_STACK_INFO]     = "=========== 线程堆栈信息 ===========",
+[PRINT_MAIN_STACK_INFO]       = "============ 主堆栈信息 ============",
+[PRINT_THREAD_STACK_OVERFLOW] = "错误：线程栈(%08x)发生溢出",
+[PRINT_MAIN_STACK_OVERFLOW]   = "错误：主栈(%08x)发生溢出",
+[PRINT_CALL_STACK_INFO]       = "查看更多函数调用栈信息，请运行：addr2line -e %s%s -a -f %.*s",
+[PRINT_CALL_STACK_ERR]        = "获取函数调用栈失败",
+[PRINT_FAULT_ON_THREAD]       =  "在线程(%s)中发生错误异常",
+[PRINT_FAULT_ON_HANDLER]      = "在中断或裸机环境下发生错误异常",
+[PRINT_REGS_TITLE]            = "========================= 寄存器信息 =========================",
+[PRINT_HFSR_VECTBL]           = "发生硬错误，原因：取中断向量时出错",
+[PRINT_MFSR_IACCVIOL]         = "发生存储器管理错误，原因：企图从不允许访问的区域取指令",
+[PRINT_MFSR_DACCVIOL]         = "发生存储器管理错误，原因：企图从不允许访问的区域读、写数据",
+[PRINT_MFSR_MUNSTKERR]        = "发生存储器管理错误，原因：出栈时企图访问不被允许的区域",
+[PRINT_MFSR_MSTKERR]          = "发生存储器管理错误，原因：入栈时企图访问不被允许的区域",
+[PRINT_MFSR_MLSPERR]          = "发生存储器管理错误，原因：惰性保存浮点状态时发生错误",
+[PRINT_BFSR_IBUSERR]          = "发生总线错误，原因：指令总线错误",
+[PRINT_BFSR_PRECISERR]        = "发生总线错误，原因：精确的数据总线错误",
+[PRINT_BFSR_IMPREISERR]       = "发生总线错误，原因：不精确的数据总线错误",
+[PRINT_BFSR_UNSTKERR]         = "发生总线错误，原因：出栈时发生错误",
+[PRINT_BFSR_STKERR]           = "发生总线错误，原因：入栈时发生错误",
+[PRINT_BFSR_LSPERR]           = "发生总线错误，原因：惰性保存浮点状态时发生错误",
+[PRINT_UFSR_UNDEFINSTR]       = "发生用法错误，原因：企图执行未定义指令",
+[PRINT_UFSR_INVSTATE]         = "发生用法错误，原因：试图切换到 ARM 状态",
+[PRINT_UFSR_INVPC]            = "发生用法错误，原因：无效的异常返回码",
+[PRINT_UFSR_NOCP]             = "发生用法错误，原因：企图执行协处理器指令",
+#if (CMB_CPU_PLATFORM_TYPE == CMB_CPU_ARM_CORTEX_M33)
+    [PRINT_UFSR_STKOF]        = "发生用法错误，原因：硬件检测到栈溢出",
+#endif
+[PRINT_UFSR_UNALIGNED]        = "发生用法错误，原因：企图执行非对齐访问",
+[PRINT_UFSR_DIVBYZERO0]       = "发生用法错误，原因：企图执行除 0 操作",
+[PRINT_DFSR_HALTED]           = "发生调试错误，原因：NVIC 停机请求",
+[PRINT_DFSR_BKPT]             = "发生调试错误，原因：执行 BKPT 指令",
+[PRINT_DFSR_DWTTRAP]          = "发生调试错误，原因：数据监测点匹配",
+[PRINT_DFSR_VCATCH]           = "发生调试错误，原因：发生向量捕获",
+[PRINT_DFSR_EXTERNAL]         = "发生调试错误，原因：外部调试请求",
+[PRINT_MMAR]                  = "发生存储器管理错误的地址：%08x",
+[PRINT_BFAR]                  = "发生总线错误的地址：%08x",

Third_Party/cm_backtrace/zh-CN/cmb_zh_CN_UTF8.h

@@ -0,0 +1,71 @@
+/*
+ * This file is part of the CmBacktrace Library.
+ *
+ * Copyright (c) 2020, Armink, <armink.ztl@gmail.com>
+ *                     Chenxuan, <chenxuan.zhao@icloud.com>
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining
+ * a copy of this software and associated documentation files (the
+ * 'Software'), to deal in the Software without restriction, including
+ * without limitation the rights to use, copy, modify, merge, publish,
+ * distribute, sublicense, and/or sell copies of the Software, and to
+ * permit persons to whom the Software is furnished to do so, subject to
+ * the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be
+ * included in all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED 'AS IS', WITHOUT WARRANTY OF ANY KIND,
+ * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.
+ * IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY
+ * CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT,
+ * TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE
+ * SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * NOTE: DO NOT include this file on the header file.
+ * Encoding: UTF-8
+ * Created on: 2020-09-06
+ */
+
+[PRINT_MAIN_STACK_CFG_ERROR]  = "错误：无法获取主栈信息，请检查主栈的相关配置",
+[PRINT_FIRMWARE_INFO]         = "固件名称：%s，硬件版本号：%s，软件版本号：%s",
+[PRINT_ASSERT_ON_THREAD]      = "在线程(%s)中发生断言",
+[PRINT_ASSERT_ON_HANDLER]     = "在中断或裸机环境下发生断言",
+[PRINT_THREAD_STACK_INFO]     = "=========== 线程堆栈信息 ===========",
+[PRINT_MAIN_STACK_INFO]       = "============ 主堆栈信息 ============",
+[PRINT_THREAD_STACK_OVERFLOW] = "错误：线程栈(%08x)发生溢出",
+[PRINT_MAIN_STACK_OVERFLOW]   = "错误：主栈(%08x)发生溢出",
+[PRINT_CALL_STACK_INFO]       = "查看更多函数调用栈信息，请运行：addr2line -e %s%s -a -f %.*s",
+[PRINT_CALL_STACK_ERR]        = "获取函数调用栈失败",
+[PRINT_FAULT_ON_THREAD]       =  "在线程(%s)中发生错误异常",
+[PRINT_FAULT_ON_HANDLER]      = "在中断或裸机环境下发生错误异常",
+[PRINT_REGS_TITLE]            = "========================= 寄存器信息 =========================",
+[PRINT_HFSR_VECTBL]           = "发生硬错误，原因：取中断向量时出错",
+[PRINT_MFSR_IACCVIOL]         = "发生存储器管理错误，原因：企图从不允许访问的区域取指令",
+[PRINT_MFSR_DACCVIOL]         = "发生存储器管理错误，原因：企图从不允许访问的区域读、写数据",
+[PRINT_MFSR_MUNSTKERR]        = "发生存储器管理错误，原因：出栈时企图访问不被允许的区域",
+[PRINT_MFSR_MSTKERR]          = "发生存储器管理错误，原因：入栈时企图访问不被允许的区域",
+[PRINT_MFSR_MLSPERR]          = "发生存储器管理错误，原因：惰性保存浮点状态时发生错误",
+[PRINT_BFSR_IBUSERR]          = "发生总线错误，原因：指令总线错误",
+[PRINT_BFSR_PRECISERR]        = "发生总线错误，原因：精确的数据总线错误",
+[PRINT_BFSR_IMPREISERR]       = "发生总线错误，原因：不精确的数据总线错误",
+[PRINT_BFSR_UNSTKERR]         = "发生总线错误，原因：出栈时发生错误",
+[PRINT_BFSR_STKERR]           = "发生总线错误，原因：入栈时发生错误",
+[PRINT_BFSR_LSPERR]           = "发生总线错误，原因：惰性保存浮点状态时发生错误",
+[PRINT_UFSR_UNDEFINSTR]       = "发生用法错误，原因：企图执行未定义指令",
+[PRINT_UFSR_INVSTATE]         = "发生用法错误，原因：试图切换到 ARM 状态",
+[PRINT_UFSR_INVPC]            = "发生用法错误，原因：无效的异常返回码",
+[PRINT_UFSR_NOCP]             = "发生用法错误，原因：企图执行协处理器指令",
+#if (CMB_CPU_PLATFORM_TYPE == CMB_CPU_ARM_CORTEX_M33)
+    [PRINT_UFSR_STKOF]        = "发生用法错误，原因：硬件检测到栈溢出",
+#endif
+[PRINT_UFSR_UNALIGNED]        = "发生用法错误，原因：企图执行非对齐访问",
+[PRINT_UFSR_DIVBYZERO0]       = "发生用法错误，原因：企图执行除 0 操作",
+[PRINT_DFSR_HALTED]           = "发生调试错误，原因：NVIC 停机请求",
+[PRINT_DFSR_BKPT]             = "发生调试错误，原因：执行 BKPT 指令",
+[PRINT_DFSR_DWTTRAP]          = "发生调试错误，原因：数据监测点匹配",
+[PRINT_DFSR_VCATCH]           = "发生调试错误，原因：发生向量捕获",
+[PRINT_DFSR_EXTERNAL]         = "发生调试错误，原因：外部调试请求",
+[PRINT_MMAR]                  = "发生存储器管理错误的地址：%08x",
+[PRINT_BFAR]                  = "发生总线错误的地址：%08x",

Third_Party/vsprintf/printf.c → Third_Party/vsprintf/new_printf.c

@@ -32,7 +32,7 @@
 
 #include <stdbool.h>
 #include <stdint.h>
-
+#include <stdio.h>
 #include "printf.h"
 
 
@@ -870,6 +870,15 @@ int printf_(const char* format, ...)
 }
 
 
+int __wrap_sprintf(char* buffer, const char* format, ...)
+{
+  va_list va;
+  va_start(va, format);
+  const int ret = _vsnprintf(_out_buffer, buffer, (size_t)-1, format, va);
+  va_end(va);
+  return ret;
+}
+
 int sprintf_(char* buffer, const char* format, ...)
 {
   va_list va;
@@ -880,6 +889,15 @@ int sprintf_(char* buffer, const char* format, ...)
 }
 
 
+int __wrap_snprintf(char* buffer, size_t count, const char* format, ...)
+{
+  va_list va;
+  va_start(va, format);
+  const int ret = _vsnprintf(_out_buffer, buffer, count, format, va);
+  va_end(va);
+  return ret;
+}
+
 int snprintf_(char* buffer, size_t count, const char* format, ...)
 {
   va_list va;
@@ -889,6 +907,11 @@ int snprintf_(char* buffer, size_t count, const char* format, ...)
   return ret;
 }
 
+int __wrap_vprintf(const char* format, va_list va)
+{
+  char buffer[1];
+  return _vsnprintf(_out_char, buffer, (size_t)-1, format, va);
+}
 
 int vprintf_(const char* format, va_list va)
 {
@@ -896,6 +919,10 @@ int vprintf_(const char* format, va_list va)
   return _vsnprintf(_out_char, buffer, (size_t)-1, format, va);
 }
 
+int __wrap_vsnprintf(char* buffer, size_t count, const char* format, va_list va)
+{
+  return _vsnprintf(_out_buffer, buffer, count, format, va);
+}
 
 int vsnprintf_(char* buffer, size_t count, const char* format, va_list va)
 {
@@ -912,3 +939,8 @@ int fctprintf(void (*out)(char character, void* arg), void* arg, const char* for
   va_end(va);
   return ret;
 }
+
+int	__wrap_fprintf (FILE *fp, const char *format, va_list va)
+{
+	return -1;
+}

Third_Party/vsprintf/printf.h

@@ -34,7 +34,7 @@
 
 #include <stdarg.h>
 #include <stddef.h>
-
+#include <stdio.h>
 
 #ifdef __cplusplus
 extern "C" {
@@ -70,7 +70,7 @@ int printf_(const char* format, ...);
  */
 //#define sprintf sprintf_
 int sprintf_(char* buffer, const char* format, ...);
-
+int __wrap_sprintf(char* buffer, const char* format, ...);
 
 /**
  * Tiny snprintf/vsnprintf implementation
@@ -86,7 +86,8 @@ int sprintf_(char* buffer, const char* format, ...);
 //#define vsnprintf vsnprintf_
 int  snprintf_(char* buffer, size_t count, const char* format, ...);
 int vsnprintf_(char* buffer, size_t count, const char* format, va_list va);
-
+int __wrap_snprintf(char* buffer, size_t count, const char* format, ...);
+int __wrap_vsnprintf(char* buffer, size_t count, const char* format, va_list va);
 
 /**
  * Tiny vprintf implementation
@@ -96,7 +97,7 @@ int vsnprintf_(char* buffer, size_t count, const char* format, va_list va);
  */
 //#define vprintf vprintf_
 int vprintf_(const char* format, va_list va);
-
+int __wrap_vprintf(const char* format, va_list va);
 
 /**
  * printf with output function
@@ -107,7 +108,7 @@ int vprintf_(const char* format, va_list va);
  * \return The number of characters that are sent to the output function, not counting the terminating null character
  */
 int fctprintf(void (*out)(char character, void* arg), void* arg, const char* format, ...);
-
+int	__wrap_fprintf (FILE *fp, const char *format, va_list va);
 
 #ifdef __cplusplus
 }

application/include/app_interface.h

@@ -28,6 +28,7 @@
 #include "core_dma.h"
 #include "core_flash.h"
 #include "core_debug.h"
+#include "core_gpio.h"
 #include "core_tick.h"
 #include "core_otp.h"
 #include "core_timer.h"

application/include/luat_conf_bsp.h

@@ -65,13 +65,16 @@
 #define LUAT_USE_FATFS 1
 #define LUAT_USE_I2CTOOLS 1
 #define LUAT_USE_SOFTKB 1
+// #define LUAT_USE_PROTOBUF
 
 #define LUAT_USE_W5500  1
 #define LUAT_USE_DHCP  1
 #define LUAT_USE_DNS  1
 #define LUAT_USE_NETWORK 1
 #define LUAT_USE_TLS 1
+// #define LUAT_USE_IOTAUTH 1
 #define LUAT_USE_LORA 1
+// #define LUAT_USE_MLX90640 1
 #define LUAT_USE_MINIZ 1
 
 //----------------------------
@@ -115,6 +118,7 @@
 #define LV_FONT_OPPOSANS_M_16
 //---------------------
 // LVGL
+#define LUAT_USE_LCD
 #define LUAT_USE_LVGL
 #define LV_DISP_DEF_REFR_PERIOD gLVFlashTime
 extern unsigned int gLVFlashTime;
@@ -173,9 +177,7 @@ extern unsigned int gLVFlashTime;
 #define LV_USE_PERF_MONITOR     1
 #define __LVGL_SLEEP_ENABLE__
 #define __LUATOS_TICK_64BIT__
-#ifndef LUAT_USE_MEMORY_OPTIMIZATION_CODE_MMAP
-#define LUAT_USE_MEMORY_OPTIMIZATION_CODE_MMAP 1
-#endif
+#define LUAT_USE_MEMORY_OPTIMIZATION_CODE_MMAP
 
 #ifndef __LV_DEBUG__
 #undef LV_USE_PERF_MONITOR

application/src/luat_base_air105.c

@@ -225,6 +225,9 @@ static const luaL_Reg loadedlibs[] = {
 #endif
 #ifdef LUAT_USE_IOTAUTH
   {"iotauth", luaopen_iotauth},
+#endif
+#ifdef LUAT_USE_PROTOBUF
+  {"protobuf", luaopen_protobuf},
 #endif
   {"usbapp", luaopen_usbapp},
   {"audio", luaopen_multimedia_audio},

application/src/luat_camera_air105.c

@@ -136,7 +136,7 @@ void DecodeQR_CBDataFun(uint8_t *Data, uint32_t Len){
 }
 
 static int32_t Camera_DrawLcd(void *DrawData, uint8_t Scan){
-    LCD_DrawStruct *draw = OS_Malloc(sizeof(LCD_DrawStruct));
+    LCD_DrawStruct *draw = malloc(sizeof(LCD_DrawStruct));
     if (!draw){
         DBG("lcd flush no memory");
         return -1;
@@ -162,10 +162,10 @@ static int32_t Camera_DrawLcd(void *DrawData, uint8_t Scan){
         draw->y2 = prvCamera.VLen + prvCamera.drawVLen -1;
         draw->Size = (draw->x2 - draw->x1 + 1) * (draw->y2 - draw->y1 + 1) * 2;
         draw->ColorMode = COLOR_MODE_RGB_565;
-        draw->Data = OS_Malloc(draw->Size);
+        draw->Data = malloc(draw->Size);
         if (!draw->Data){
             DBG("lcd flush data no memory");
-            OS_Free(draw);
+            free(draw);
             return -1;
         }
         memcpy(draw->Data, DrawData, draw->Size);
@@ -181,7 +181,7 @@ static int32_t Camera_DrawLcd(void *DrawData, uint8_t Scan){
         Core_CameraDraw(draw);
     }else{
         DBG("Scan error");
-        OS_Free(draw);
+        free(draw);
         return -1;
     }
 }
@@ -398,7 +398,7 @@ int luat_camera_capture(int id, uint8_t quality, const char *path)
 
 static voidpf luat_zip_zalloc(    voidpf opaque, uint32_t items, uint32_t size)
 {
-	voidpf *p = malloc(items * size);
+	voidpf *p = zalloc(items * size);
 	return p;
 }
 
@@ -434,10 +434,10 @@ static int luat_camera_video_zip(void *data ,void *param)
 		}
     }
 
-#undef zalloc
+
     strm.zalloc = luat_zip_zalloc;
     strm.zfree = luat_zip_free;
-#define zalloc OS_Zalloc
+
     ret = deflateInit2_(&strm, 1, Z_DEFLATED, 12, 6,
     		Z_DEFAULT_STRATEGY, ZLIB_VERSION, (int)sizeof(z_stream));
     if (Z_OK == ret)

application/src/luat_i2c_air105.c

@@ -33,12 +33,15 @@ int luat_i2c_exist(int id) {
 }
     
 int luat_i2c_setup(int id, int speed, int slaveaddr) {
-    if (speed == 0)
+    if (speed == 0) {
         speed = 100 * 1000; // SLOW
-    else if (speed == 1)
+    }
+    else if (speed == 1) {
         speed = 400 * 1000; // FAST
-    else if (speed == 2)
+    }
+    else if (speed == 2) {
         speed = 400 * 1000; // SuperFast
+    }
 	GPIO_Iomux(GPIOE_06, 2);
 	GPIO_Iomux(GPIOE_07, 2);
 	I2C_MasterSetup(id, speed);

application/src/luat_malloc_air105.c

@@ -41,23 +41,23 @@ static uint64_t luavm_pool_data[25 * 1024];
 //  管理系统内存
 
 void* luat_heap_malloc(size_t len) {
-    return OS_Malloc(len);
+    return malloc(len);
 }
 
 void luat_heap_free(void* ptr) {
-    OS_Free(ptr);
+    free(ptr);
 }
 
 void* luat_heap_realloc(void* ptr, size_t len) {
-    return OS_Realloc(ptr, len);
+    return realloc(ptr, len);
 }
 
 void* luat_heap_calloc(size_t count, size_t _size) {
-    return OS_Calloc(count,_size);
+    return calloc(count,_size);
 }
 
 void* luat_heap_zalloc(size_t _size) {
-    return OS_Zalloc(_size);
+    return zalloc(_size);
 }
 
 //------------------------------------------------
@@ -88,13 +88,13 @@ void* luat_heap_alloc(void *ud, void *ptr, size_t osize, size_t nsize) {
 //            else {
 //                // 释放内存块
 //                LLOGD("free %p ", ptr);
-//                OS_Free(ptr);
+//                free(ptr);
 //                return NULL;
 //            }
 //        }
 //        else {
 //            // 申请内存块
-//            ptr = OS_Malloc(nsize);
+//            ptr = malloc(nsize);
 //            LLOGD("malloc %p type=%d size=%d", ptr, osize, nsize);
 //            return ptr;
 //        }

application/src/luat_spi_air105.c

@@ -230,7 +230,7 @@ int luat_lcd_draw_no_block(luat_lcd_conf_t* conf, uint16_t x1, uint16_t y1, uint
 		{
 			LLOGD("lcd flush delay %ums, status %u,%u,%u,%d", retry_cnt, cache_len, x2-x1+1, y2-y1+1, last_flush);
 		}
-		LCD_DrawStruct *draw = OS_Zalloc(sizeof(LCD_DrawStruct));
+		LCD_DrawStruct *draw = zalloc(sizeof(LCD_DrawStruct));
 		if (!draw)
 		{
 			LLOGE("lcd flush no memory");
@@ -256,11 +256,11 @@ int luat_lcd_draw_no_block(luat_lcd_conf_t* conf, uint16_t x1, uint16_t y1, uint
 	    draw->yoffset = conf->yoffset;
 	    draw->Size = (draw->x2 - draw->x1 + 1) * (draw->y2 - draw->y1 + 1) * 2;
 	    draw->ColorMode = COLOR_MODE_RGB_565;
-	    draw->Data = OS_Malloc(draw->Size);
+	    draw->Data = malloc(draw->Size);
 		if (!draw->Data)
 		{
 			LLOGE("lcd flush data no memory");
-			OS_Free(draw);
+			free(draw);
 			return -1;
 		}
 	    memcpy(draw->Data, color, draw->Size);

bsp/air105/hal/core_dcmi.c

@@ -177,7 +177,7 @@ void DCMI_CaptureSwitch(uint8_t OnOff, uint32_t BufLen, uint32_t ImageW, uint32_
 		}
 		for(i = 0; i < DCMI_RXBUF_BAND; i++)
 		{
-			prvDCMI.uBuf[i].pu32 = OS_Malloc(BufLen * sizeof(uint32_t));
+			prvDCMI.uBuf[i].pu32 = malloc(BufLen * sizeof(uint32_t));
 		}
 		prvDCMI.RxDMASn = 0;
 		prvDCMI.BufLen = BufLen;
@@ -200,7 +200,7 @@ void DCMI_CaptureSwitch(uint8_t OnOff, uint32_t BufLen, uint32_t ImageW, uint32_
 		{
 			if (prvDCMI.uBuf[i].pu32)
 			{
-				OS_Free(prvDCMI.uBuf[i].pu32);
+				free(prvDCMI.uBuf[i].pu32);
 				prvDCMI.uBuf[i].pu32 = 0;
 			}
 		}

bsp/air105/hal/core_debug.c

@@ -345,7 +345,7 @@ void DBG_HexPrintf(void *Data, unsigned int len)
     {
     	return ;
     }
-    uart_buf = OS_Zalloc(len * 3 + 2);
+    uart_buf = zalloc(len * 3 + 2);
 
     if (!uart_buf)
     {
@@ -361,7 +361,55 @@ void DBG_HexPrintf(void *Data, unsigned int len)
     uart_buf[j++] = '\n';
 
     prvDBGCtrl.TxFun(uart_buf, len * 3 + 2);
-	OS_Free(uart_buf);
+	free(uart_buf);
+}
+
+int	__wrap_putc (int ch, FILE * fd)
+{
+	return -1;
+}
+
+int	__wrap_putchar (int ch)
+{
+	char c = ch;
+	add_printf_data(&c, 1);
+	return 1;
+}
+int	__wrap_puts (const char *buf)
+{
+	int len = strlen(buf);
+	add_printf_data(buf, len);
+	return len;
+}
+
+void __wrap_printf(const char* format, ...)
+{
+	char *buf = NULL;
+	char isr_buf[128];
+	int len;
+	va_list ap;
+	if (!prvDBGCtrl.AppMode) return;
+	va_start(ap, format);
+	if (OS_CheckInIrq())
+	{
+		buf = isr_buf;
+		len = vsnprintf(buf, 127, format, ap);
+	}
+	else
+	{
+		buf = zalloc(1024);
+		len = vsnprintf(buf, 1023, format, ap);
+	}
+	va_end(ap);
+#ifndef __DEBUG__
+    prvDBGCtrl.TxFun( buf, len);
+#else
+    add_printf_data(buf, len);
+#endif
+	if (!OS_CheckInIrq())
+	{
+		free(buf);
+	}
 }
 
 void DBG_Printf(const char* format, ...)
@@ -375,12 +423,12 @@ void DBG_Printf(const char* format, ...)
 	if (OS_CheckInIrq())
 	{
 		buf = isr_buf;
-		len = vsnprintf_(buf, 127, format, ap);
+		len = vsnprintf(buf, 127, format, ap);
 	}
 	else
 	{
-		buf = OS_Zalloc(1024);
-		len = vsnprintf_(buf, 1023, format, ap);
+		buf = zalloc(1024);
+		len = vsnprintf(buf, 1023, format, ap);
 	}
 	va_end(ap);
 #ifndef __DEBUG__
@@ -390,7 +438,7 @@ void DBG_Printf(const char* format, ...)
 #endif
 	if (!OS_CheckInIrq())
 	{
-		OS_Free(buf);
+		free(buf);
 	}
 }
 

bsp/air105/hal/core_hwtimer.c

@@ -615,10 +615,10 @@ void HWTimer_InitOperationQueue(uint8_t HWTimerID, uint32_t nCount, uint32_t Rep
 	prvHWTimer[HWTimerID].TotalRepeat = Repeat;
 	if (prvHWTimer[HWTimerID].Cmd)
 	{
-		OS_Free(prvHWTimer[HWTimerID].Cmd);
+		free(prvHWTimer[HWTimerID].Cmd);
 		prvHWTimer[HWTimerID].Cmd = NULL;
 	}
-	prvHWTimer[HWTimerID].Cmd = OS_Zalloc((nCount + 1) * sizeof(OPQueue_CmdStruct));
+	prvHWTimer[HWTimerID].Cmd = zalloc((nCount + 1) * sizeof(OPQueue_CmdStruct));
 	prvHWTimer[HWTimerID].CmdQueuePos = 0;
 	if (CmdDoneCB)
 	{
@@ -680,7 +680,7 @@ void HWTimer_ClearOperationQueue(uint8_t HWTimerID)
 
 void HWTimer_FreeOperationQueue(uint8_t HWTimerID)
 {
-	OS_Free(prvHWTimer[HWTimerID].Cmd);
+	free(prvHWTimer[HWTimerID].Cmd);
 	prvHWTimer[HWTimerID].Cmd = NULL;
 	prvHWTimer[HWTimerID].CmdDoneCB = prvHWTimer_DummyCB;
 }

bsp/air105/hal/core_timer.c

@@ -223,7 +223,7 @@ void Timer_WakeupRun(void)
 Timer_t * Timer_Create(CBFuncEx_t CB, void *Param, void *NotifyTask)
 {
 #ifdef __BUILD_OS__
-	Timer_t *Timer = OS_Malloc(sizeof(Timer_t));
+	Timer_t *Timer = malloc(sizeof(Timer_t));
 	if (Timer)
 	{
 		Timer_Setting(Timer, CB, Param, NotifyTask);
@@ -313,7 +313,7 @@ void Timer_Release(Timer_t *Timer)
 	Timer_Stop(Timer);
 	if ((uint32_t)Timer >= (uint32_t)(&__os_heap_start))
 	{
-		OS_Free(Timer);
+		free(Timer);
 	}
 }
 

bsp/air105/include/app_inc.h

@@ -57,10 +57,6 @@ typedef long long           int64_t;
 #include "core_task.h"
 #include "core_service.h"
 
-#define mallc OS_Malloc
-#define zallc OS_Zalloc
-#define reallc OS_Realloc
-#define free OS_Free
 #define IRQ_LOWEST_PRIORITY	configLIBRARY_LOWEST_INTERRUPT_PRIORITY
 #define IRQ_MAX_PRIORITY	configLIBRARY_MAX_SYSCALL_INTERRUPT_PRIORITY
 #endif

bsp/air105/platform/app_main.c

@@ -48,8 +48,22 @@ const uint32_t __attribute__((section (".app_info")))
 
 static void prvSystemHaltResetCtrl(void)
 {
+	if (WDT->WDT_STAT)
+	{
+		DBG_Trace("触发看门狗中断");
+#ifdef __BUILD_OS__
+		Core_PrintServiceStack();
+#endif
+	}
 #ifdef __DEBUG__
-	WDT_Feed();
+	DBG_Trace("当前处于调试模式，将处于死循环状态");
+	while (1)
+	{
+		WDT_Feed();
+	}
+#else
+	WDT_ModeConfig(WDT_Mode_CPUReset);
+	DBG_Trace("当前处于工作模式，将在下次看门狗中断到来后重启");
 #endif
 }
 void SystemInit(void)

bsp/air105/platform/bl_main.c

@@ -646,6 +646,7 @@ int main(void)
 	DBG_Init(0);
 	FileSystem_Init();
 	OS_InitBuffer(&prvBL.FWDataBuffer, SPI_FLASH_BLOCK_SIZE);
+
 LOCAL_UPGRADE_START:
 	Local_Upgrade();
 	Uart_BaseInit(DBG_UART_ID, DBG_UART_BR, 0, UART_DATA_BIT8, UART_PARITY_NONE, UART_STOP_BIT1, NULL);
@@ -658,9 +659,9 @@ LOCAL_UPGRADE_START:
 	if (__APP_START_MAGIC__ == AppInfo[0])
 	{
 #ifdef __DEBUG__
-		DBG_INFO("bootloader build debug %s %s %x!", __DATE__, __TIME__, QSPI->DEVICE_PARA);
+		DBG_INFO("bootloader build debug %s %s!", __DATE__, __TIME__);
 #else
-		DBG_INFO("bootloader build release %s %s!", __DATE__, __TIME__, QSPI->DEVICE_PARA);
+		DBG_INFO("bootloader build release %s %s!", __DATE__, __TIME__);
 #endif
 		Jump_AppRun(AppInfo[1]);
 	}

bsp/air105/test/test_camera.c

@@ -685,11 +685,11 @@ static int32_t prvCamera_DCMICB(void *pData, void *pParam)
 	draw->yoffset = 0;
 	draw->Size = (draw->x2 - draw->x1 + 1) * (draw->y2 - draw->y1 + 1) * 2;
 	draw->ColorMode = COLOR_MODE_RGB_565;
-	draw->Data = OS_Malloc(draw->Size);
+	draw->Data = malloc(draw->Size);
 	if (!draw->Data)
 	{
 		DBG("lcd flush data no memory");
-		OS_Free(draw);
+		free(draw);
 		return -1;
 	}
 //	DBG_HexPrintf(RxBuf->Data, 8);

bsp/air105/test/test_spi.c

@@ -51,8 +51,8 @@ void SPI_TestInit(uint8_t SpiID, uint32_t Speed, uint32_t TestLen)
 	SPI_MasterInit(SpiID, 8, SPI_MODE_0, Speed, prvSPI_CB, NULL);
 	SPI_DMATxInit(SpiID, DMA1_STREAM_6, 0);
 	SPI_DMARxInit(SpiID, DMA1_STREAM_7, 0);
-	prvTxBuf = OS_Malloc(TestLen);
-	prvRxBuf = OS_Malloc(TestLen);
+	prvTxBuf = malloc(TestLen);
+	prvRxBuf = malloc(TestLen);
 }
 
 void SPI_TestOnce(uint8_t SpiID, uint32_t TestLen)

bsp/cmsis/include/dsp/basic_math_functions.h

@@ -0,0 +1,880 @@
+/******************************************************************************
+ * @file     basic_math_functions.h
+ * @brief    Public header file for CMSIS DSP Library
+ * @version  V1.10.0
+ * @date     08 July 2021
+ * Target Processor: Cortex-M and Cortex-A cores
+ ******************************************************************************/
+/*
+ * Copyright (c) 2010-2020 Arm Limited or its affiliates. All rights reserved.
+ *
+ * SPDX-License-Identifier: Apache-2.0
+ *
+ * Licensed under the Apache License, Version 2.0 (the License); you may
+ * not use this file except in compliance with the License.
+ * You may obtain a copy of the License at
+ *
+ * www.apache.org/licenses/LICENSE-2.0
+ *
+ * Unless required by applicable law or agreed to in writing, software
+ * distributed under the License is distributed on an AS IS BASIS, WITHOUT
+ * WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ * See the License for the specific language governing permissions and
+ * limitations under the License.
+ */
+
+ 
+#ifndef _BASIC_MATH_FUNCTIONS_H_
+#define _BASIC_MATH_FUNCTIONS_H_
+
+#include "arm_math_types.h"
+#include "arm_math_memory.h"
+
+#include "dsp/none.h"
+#include "dsp/utils.h"
+
+
+#ifdef   __cplusplus
+extern "C"
+{
+#endif
+
+/**
+ * @defgroup groupMath Basic Math Functions
+ */
+
+ /**
+   * @brief Q7 vector multiplication.
+   * @param[in]  pSrcA      points to the first input vector
+   * @param[in]  pSrcB      points to the second input vector
+   * @param[out] pDst       points to the output vector
+   * @param[in]  blockSize  number of samples in each vector
+   */
+  void arm_mult_q7(
+  const q7_t * pSrcA,
+  const q7_t * pSrcB,
+        q7_t * pDst,
+        uint32_t blockSize);
+
+
+  /**
+   * @brief Q15 vector multiplication.
+   * @param[in]  pSrcA      points to the first input vector
+   * @param[in]  pSrcB      points to the second input vector
+   * @param[out] pDst       points to the output vector
+   * @param[in]  blockSize  number of samples in each vector
+   */
+  void arm_mult_q15(
+  const q15_t * pSrcA,
+  const q15_t * pSrcB,
+        q15_t * pDst,
+        uint32_t blockSize);
+
+
+  /**
+   * @brief Q31 vector multiplication.
+   * @param[in]  pSrcA      points to the first input vector
+   * @param[in]  pSrcB      points to the second input vector
+   * @param[out] pDst       points to the output vector
+   * @param[in]  blockSize  number of samples in each vector
+   */
+  void arm_mult_q31(
+  const q31_t * pSrcA,
+  const q31_t * pSrcB,
+        q31_t * pDst,
+        uint32_t blockSize);
+
+
+  /**
+   * @brief Floating-point vector multiplication.
+   * @param[in]  pSrcA      points to the first input vector
+   * @param[in]  pSrcB      points to the second input vector
+   * @param[out] pDst       points to the output vector
+   * @param[in]  blockSize  number of samples in each vector
+   */
+  void arm_mult_f32(
+  const float32_t * pSrcA,
+  const float32_t * pSrcB,
+        float32_t * pDst,
+        uint32_t blockSize);
+
+
+
+/**
+ * @brief Floating-point vector multiplication.
+ * @param[in]  pSrcA      points to the first input vector
+ * @param[in]  pSrcB      points to the second input vector
+ * @param[out] pDst       points to the output vector
+ * @param[in]  blockSize  number of samples in each vector
+ */
+void arm_mult_f64(
+const float64_t * pSrcA,
+const float64_t * pSrcB,
+	  float64_t * pDst,
+	  uint32_t blockSize);
+
+
+
+ /**
+   * @brief Floating-point vector addition.
+   * @param[in]  pSrcA      points to the first input vector
+   * @param[in]  pSrcB      points to the second input vector
+   * @param[out] pDst       points to the output vector
+   * @param[in]  blockSize  number of samples in each vector
+   */
+  void arm_add_f32(
+  const float32_t * pSrcA,
+  const float32_t * pSrcB,
+        float32_t * pDst,
+        uint32_t blockSize);
+
+
+
+/**
+  * @brief Floating-point vector addition.
+  * @param[in]  pSrcA      points to the first input vector
+  * @param[in]  pSrcB      points to the second input vector
+  * @param[out] pDst       points to the output vector
+  * @param[in]  blockSize  number of samples in each vector
+  */
+ void arm_add_f64(
+ const float64_t * pSrcA,
+ const float64_t * pSrcB,
+	   float64_t * pDst,
+	   uint32_t blockSize);
+
+
+
+  /**
+   * @brief Q7 vector addition.
+   * @param[in]  pSrcA      points to the first input vector
+   * @param[in]  pSrcB      points to the second input vector
+   * @param[out] pDst       points to the output vector
+   * @param[in]  blockSize  number of samples in each vector
+   */
+  void arm_add_q7(
+  const q7_t * pSrcA,
+  const q7_t * pSrcB,
+        q7_t * pDst,
+        uint32_t blockSize);
+
+
+  /**
+   * @brief Q15 vector addition.
+   * @param[in]  pSrcA      points to the first input vector
+   * @param[in]  pSrcB      points to the second input vector
+   * @param[out] pDst       points to the output vector
+   * @param[in]  blockSize  number of samples in each vector
+   */
+  void arm_add_q15(
+  const q15_t * pSrcA,
+  const q15_t * pSrcB,
+        q15_t * pDst,
+        uint32_t blockSize);
+
+
+  /**
+   * @brief Q31 vector addition.
+   * @param[in]  pSrcA      points to the first input vector
+   * @param[in]  pSrcB      points to the second input vector
+   * @param[out] pDst       points to the output vector
+   * @param[in]  blockSize  number of samples in each vector
+   */
+  void arm_add_q31(
+  const q31_t * pSrcA,
+  const q31_t * pSrcB,
+        q31_t * pDst,
+        uint32_t blockSize);
+
+
+  /**
+   * @brief Floating-point vector subtraction.
+   * @param[in]  pSrcA      points to the first input vector
+   * @param[in]  pSrcB      points to the second input vector
+   * @param[out] pDst       points to the output vector
+   * @param[in]  blockSize  number of samples in each vector
+   */
+  void arm_sub_f32(
+  const float32_t * pSrcA,
+  const float32_t * pSrcB,
+        float32_t * pDst,
+        uint32_t blockSize);
+
+
+
+  /**
+   * @brief Floating-point vector subtraction.
+   * @param[in]  pSrcA      points to the first input vector
+   * @param[in]  pSrcB      points to the second input vector
+   * @param[out] pDst       points to the output vector
+   * @param[in]  blockSize  number of samples in each vector
+   */
+  void arm_sub_f64(
+  const float64_t * pSrcA,
+  const float64_t * pSrcB,
+        float64_t * pDst,
+        uint32_t blockSize);
+
+
+
+  /**
+   * @brief Q7 vector subtraction.
+   * @param[in]  pSrcA      points to the first input vector
+   * @param[in]  pSrcB      points to the second input vector
+   * @param[out] pDst       points to the output vector
+   * @param[in]  blockSize  number of samples in each vector
+   */
+  void arm_sub_q7(
+  const q7_t * pSrcA,
+  const q7_t * pSrcB,
+        q7_t * pDst,
+        uint32_t blockSize);
+
+
+  /**
+   * @brief Q15 vector subtraction.
+   * @param[in]  pSrcA      points to the first input vector
+   * @param[in]  pSrcB      points to the second input vector
+   * @param[out] pDst       points to the output vector
+   * @param[in]  blockSize  number of samples in each vector
+   */
+  void arm_sub_q15(
+  const q15_t * pSrcA,
+  const q15_t * pSrcB,
+        q15_t * pDst,
+        uint32_t blockSize);
+
+
+  /**
+   * @brief Q31 vector subtraction.
+   * @param[in]  pSrcA      points to the first input vector
+   * @param[in]  pSrcB      points to the second input vector
+   * @param[out] pDst       points to the output vector
+   * @param[in]  blockSize  number of samples in each vector
+   */
+  void arm_sub_q31(
+  const q31_t * pSrcA,
+  const q31_t * pSrcB,
+        q31_t * pDst,
+        uint32_t blockSize);
+
+
+  /**
+   * @brief Multiplies a floating-point vector by a scalar.
+   * @param[in]  pSrc       points to the input vector
+   * @param[in]  scale      scale factor to be applied
+   * @param[out] pDst       points to the output vector
+   * @param[in]  blockSize  number of samples in the vector
+   */
+  void arm_scale_f32(
+  const float32_t * pSrc,
+        float32_t scale,
+        float32_t * pDst,
+        uint32_t blockSize);
+
+
+
+  /**
+   * @brief Multiplies a floating-point vector by a scalar.
+   * @param[in]  pSrc       points to the input vector
+   * @param[in]  scale      scale factor to be applied
+   * @param[out] pDst       points to the output vector
+   * @param[in]  blockSize  number of samples in the vector
+   */
+  void arm_scale_f64(
+  const float64_t * pSrc,
+        float64_t scale,
+        float64_t * pDst,
+        uint32_t blockSize);
+
+
+
+  /**
+   * @brief Multiplies a Q7 vector by a scalar.
+   * @param[in]  pSrc        points to the input vector
+   * @param[in]  scaleFract  fractional portion of the scale value
+   * @param[in]  shift       number of bits to shift the result by
+   * @param[out] pDst        points to the output vector
+   * @param[in]  blockSize   number of samples in the vector
+   */
+  void arm_scale_q7(
+  const q7_t * pSrc,
+        q7_t scaleFract,
+        int8_t shift,
+        q7_t * pDst,
+        uint32_t blockSize);
+
+
+  /**
+   * @brief Multiplies a Q15 vector by a scalar.
+   * @param[in]  pSrc        points to the input vector
+   * @param[in]  scaleFract  fractional portion of the scale value
+   * @param[in]  shift       number of bits to shift the result by
+   * @param[out] pDst        points to the output vector
+   * @param[in]  blockSize   number of samples in the vector
+   */
+  void arm_scale_q15(
+  const q15_t * pSrc,
+        q15_t scaleFract,
+        int8_t shift,
+        q15_t * pDst,
+        uint32_t blockSize);
+
+
+  /**
+   * @brief Multiplies a Q31 vector by a scalar.
+   * @param[in]  pSrc        points to the input vector
+   * @param[in]  scaleFract  fractional portion of the scale value
+   * @param[in]  shift       number of bits to shift the result by
+   * @param[out] pDst        points to the output vector
+   * @param[in]  blockSize   number of samples in the vector
+   */
+  void arm_scale_q31(
+  const q31_t * pSrc,
+        q31_t scaleFract,
+        int8_t shift,
+        q31_t * pDst,
+        uint32_t blockSize);
+
+
+  /**
+   * @brief Q7 vector absolute value.
+   * @param[in]  pSrc       points to the input buffer
+   * @param[out] pDst       points to the output buffer
+   * @param[in]  blockSize  number of samples in each vector
+   */
+  void arm_abs_q7(
+  const q7_t * pSrc,
+        q7_t * pDst,
+        uint32_t blockSize);
+
+
+  /**
+   * @brief Floating-point vector absolute value.
+   * @param[in]  pSrc       points to the input buffer
+   * @param[out] pDst       points to the output buffer
+   * @param[in]  blockSize  number of samples in each vector
+   */
+  void arm_abs_f32(
+  const float32_t * pSrc,
+        float32_t * pDst,
+        uint32_t blockSize);
+
+
+
+/**
+ * @brief Floating-point vector absolute value.
+ * @param[in]  pSrc       points to the input buffer
+ * @param[out] pDst       points to the output buffer
+ * @param[in]  blockSize  number of samples in each vector
+ */
+void arm_abs_f64(
+const float64_t * pSrc,
+	  float64_t * pDst,
+	  uint32_t blockSize);
+
+
+
+  /**
+   * @brief Q15 vector absolute value.
+   * @param[in]  pSrc       points to the input buffer
+   * @param[out] pDst       points to the output buffer
+   * @param[in]  blockSize  number of samples in each vector
+   */
+  void arm_abs_q15(
+  const q15_t * pSrc,
+        q15_t * pDst,
+        uint32_t blockSize);
+
+
+  /**
+   * @brief Q31 vector absolute value.
+   * @param[in]  pSrc       points to the input buffer
+   * @param[out] pDst       points to the output buffer
+   * @param[in]  blockSize  number of samples in each vector
+   */
+  void arm_abs_q31(
+  const q31_t * pSrc,
+        q31_t * pDst,
+        uint32_t blockSize);
+
+
+  /**
+   * @brief Dot product of floating-point vectors.
+   * @param[in]  pSrcA      points to the first input vector
+   * @param[in]  pSrcB      points to the second input vector
+   * @param[in]  blockSize  number of samples in each vector
+   * @param[out] result     output result returned here
+   */
+  void arm_dot_prod_f32(
+  const float32_t * pSrcA,
+  const float32_t * pSrcB,
+        uint32_t blockSize,
+        float32_t * result);
+
+
+
+/**
+ * @brief Dot product of floating-point vectors.
+ * @param[in]  pSrcA      points to the first input vector
+ * @param[in]  pSrcB      points to the second input vector
+ * @param[in]  blockSize  number of samples in each vector
+ * @param[out] result     output result returned here
+ */
+void arm_dot_prod_f64(
+const float64_t * pSrcA,
+const float64_t * pSrcB,
+	  uint32_t blockSize,
+	  float64_t * result);
+
+
+
+  /**
+   * @brief Dot product of Q7 vectors.
+   * @param[in]  pSrcA      points to the first input vector
+   * @param[in]  pSrcB      points to the second input vector
+   * @param[in]  blockSize  number of samples in each vector
+   * @param[out] result     output result returned here
+   */
+  void arm_dot_prod_q7(
+  const q7_t * pSrcA,
+  const q7_t * pSrcB,
+        uint32_t blockSize,
+        q31_t * result);
+
+
+  /**
+   * @brief Dot product of Q15 vectors.
+   * @param[in]  pSrcA      points to the first input vector
+   * @param[in]  pSrcB      points to the second input vector
+   * @param[in]  blockSize  number of samples in each vector
+   * @param[out] result     output result returned here
+   */
+  void arm_dot_prod_q15(
+  const q15_t * pSrcA,
+  const q15_t * pSrcB,
+        uint32_t blockSize,
+        q63_t * result);
+
+
+  /**
+   * @brief Dot product of Q31 vectors.
+   * @param[in]  pSrcA      points to the first input vector
+   * @param[in]  pSrcB      points to the second input vector
+   * @param[in]  blockSize  number of samples in each vector
+   * @param[out] result     output result returned here
+   */
+  void arm_dot_prod_q31(
+  const q31_t * pSrcA,
+  const q31_t * pSrcB,
+        uint32_t blockSize,
+        q63_t * result);
+
+
+  /**
+   * @brief  Shifts the elements of a Q7 vector a specified number of bits.
+   * @param[in]  pSrc       points to the input vector
+   * @param[in]  shiftBits  number of bits to shift.  A positive value shifts left; a negative value shifts right.
+   * @param[out] pDst       points to the output vector
+   * @param[in]  blockSize  number of samples in the vector
+   */
+  void arm_shift_q7(
+  const q7_t * pSrc,
+        int8_t shiftBits,
+        q7_t * pDst,
+        uint32_t blockSize);
+
+
+  /**
+   * @brief  Shifts the elements of a Q15 vector a specified number of bits.
+   * @param[in]  pSrc       points to the input vector
+   * @param[in]  shiftBits  number of bits to shift.  A positive value shifts left; a negative value shifts right.
+   * @param[out] pDst       points to the output vector
+   * @param[in]  blockSize  number of samples in the vector
+   */
+  void arm_shift_q15(
+  const q15_t * pSrc,
+        int8_t shiftBits,
+        q15_t * pDst,
+        uint32_t blockSize);
+
+
+  /**
+   * @brief  Shifts the elements of a Q31 vector a specified number of bits.
+   * @param[in]  pSrc       points to the input vector
+   * @param[in]  shiftBits  number of bits to shift.  A positive value shifts left; a negative value shifts right.
+   * @param[out] pDst       points to the output vector
+   * @param[in]  blockSize  number of samples in the vector
+   */
+  void arm_shift_q31(
+  const q31_t * pSrc,
+        int8_t shiftBits,
+        q31_t * pDst,
+        uint32_t blockSize);
+
+
+/**
+ * @brief  Adds a constant offset to a floating-point vector.
+ * @param[in]  pSrc       points to the input vector
+ * @param[in]  offset     is the offset to be added
+ * @param[out] pDst       points to the output vector
+ * @param[in]  blockSize  number of samples in the vector
+ */
+void arm_offset_f64(
+const float64_t * pSrc,
+	  float64_t offset,
+	  float64_t * pDst,
+	  uint32_t blockSize);
+
+
+
+  /**
+   * @brief  Adds a constant offset to a floating-point vector.
+   * @param[in]  pSrc       points to the input vector
+   * @param[in]  offset     is the offset to be added
+   * @param[out] pDst       points to the output vector
+   * @param[in]  blockSize  number of samples in the vector
+   */
+  void arm_offset_f32(
+  const float32_t * pSrc,
+        float32_t offset,
+        float32_t * pDst,
+        uint32_t blockSize);
+
+
+
+  /**
+   * @brief  Adds a constant offset to a Q7 vector.
+   * @param[in]  pSrc       points to the input vector
+   * @param[in]  offset     is the offset to be added
+   * @param[out] pDst       points to the output vector
+   * @param[in]  blockSize  number of samples in the vector
+   */
+  void arm_offset_q7(
+  const q7_t * pSrc,
+        q7_t offset,
+        q7_t * pDst,
+        uint32_t blockSize);
+
+
+  /**
+   * @brief  Adds a constant offset to a Q15 vector.
+   * @param[in]  pSrc       points to the input vector
+   * @param[in]  offset     is the offset to be added
+   * @param[out] pDst       points to the output vector
+   * @param[in]  blockSize  number of samples in the vector
+   */
+  void arm_offset_q15(
+  const q15_t * pSrc,
+        q15_t offset,
+        q15_t * pDst,
+        uint32_t blockSize);
+
+
+  /**
+   * @brief  Adds a constant offset to a Q31 vector.
+   * @param[in]  pSrc       points to the input vector
+   * @param[in]  offset     is the offset to be added
+   * @param[out] pDst       points to the output vector
+   * @param[in]  blockSize  number of samples in the vector
+   */
+  void arm_offset_q31(
+  const q31_t * pSrc,
+        q31_t offset,
+        q31_t * pDst,
+        uint32_t blockSize);
+
+
+  /**
+   * @brief  Negates the elements of a floating-point vector.
+   * @param[in]  pSrc       points to the input vector
+   * @param[out] pDst       points to the output vector
+   * @param[in]  blockSize  number of samples in the vector
+   */
+  void arm_negate_f32(
+  const float32_t * pSrc,
+        float32_t * pDst,
+        uint32_t blockSize);
+
+
+
+/**
+ * @brief  Negates the elements of a floating-point vector.
+ * @param[in]  pSrc       points to the input vector
+ * @param[out] pDst       points to the output vector
+ * @param[in]  blockSize  number of samples in the vector
+ */
+void arm_negate_f64(
+const float64_t * pSrc,
+	  float64_t * pDst,
+	  uint32_t blockSize);
+
+
+
+  /**
+   * @brief  Negates the elements of a Q7 vector.
+   * @param[in]  pSrc       points to the input vector
+   * @param[out] pDst       points to the output vector
+   * @param[in]  blockSize  number of samples in the vector
+   */
+  void arm_negate_q7(
+  const q7_t * pSrc,
+        q7_t * pDst,
+        uint32_t blockSize);
+
+
+  /**
+   * @brief  Negates the elements of a Q15 vector.
+   * @param[in]  pSrc       points to the input vector
+   * @param[out] pDst       points to the output vector
+   * @param[in]  blockSize  number of samples in the vector
+   */
+  void arm_negate_q15(
+  const q15_t * pSrc,
+        q15_t * pDst,
+        uint32_t blockSize);
+
+
+  /**
+   * @brief  Negates the elements of a Q31 vector.
+   * @param[in]  pSrc       points to the input vector
+   * @param[out] pDst       points to the output vector
+   * @param[in]  blockSize  number of samples in the vector
+   */
+  void arm_negate_q31(
+  const q31_t * pSrc,
+        q31_t * pDst,
+        uint32_t blockSize);
+
+/**
+   * @brief         Compute the logical bitwise AND of two fixed-point vectors.
+   * @param[in]     pSrcA      points to input vector A
+   * @param[in]     pSrcB      points to input vector B
+   * @param[out]    pDst       points to output vector
+   * @param[in]     blockSize  number of samples in each vector
+   * @return        none
+   */
+  void arm_and_u16(
+    const uint16_t * pSrcA,
+    const uint16_t * pSrcB,
+          uint16_t * pDst,
+          uint32_t blockSize);
+
+  /**
+   * @brief         Compute the logical bitwise AND of two fixed-point vectors.
+   * @param[in]     pSrcA      points to input vector A
+   * @param[in]     pSrcB      points to input vector B
+   * @param[out]    pDst       points to output vector
+   * @param[in]     blockSize  number of samples in each vector
+   * @return        none
+   */
+  void arm_and_u32(
+    const uint32_t * pSrcA,
+    const uint32_t * pSrcB,
+          uint32_t * pDst,
+          uint32_t blockSize);
+
+  /**
+   * @brief         Compute the logical bitwise AND of two fixed-point vectors.
+   * @param[in]     pSrcA      points to input vector A
+   * @param[in]     pSrcB      points to input vector B
+   * @param[out]    pDst       points to output vector
+   * @param[in]     blockSize  number of samples in each vector
+   * @return        none
+   */
+  void arm_and_u8(
+    const uint8_t * pSrcA,
+    const uint8_t * pSrcB,
+          uint8_t * pDst,
+          uint32_t blockSize);
+
+  /**
+   * @brief         Compute the logical bitwise OR of two fixed-point vectors.
+   * @param[in]     pSrcA      points to input vector A
+   * @param[in]     pSrcB      points to input vector B
+   * @param[out]    pDst       points to output vector
+   * @param[in]     blockSize  number of samples in each vector
+   * @return        none
+   */
+  void arm_or_u16(
+    const uint16_t * pSrcA,
+    const uint16_t * pSrcB,
+          uint16_t * pDst,
+          uint32_t blockSize);
+
+  /**
+   * @brief         Compute the logical bitwise OR of two fixed-point vectors.
+   * @param[in]     pSrcA      points to input vector A
+   * @param[in]     pSrcB      points to input vector B
+   * @param[out]    pDst       points to output vector
+   * @param[in]     blockSize  number of samples in each vector
+   * @return        none
+   */
+  void arm_or_u32(
+    const uint32_t * pSrcA,
+    const uint32_t * pSrcB,
+          uint32_t * pDst,
+          uint32_t blockSize);
+
+  /**
+   * @brief         Compute the logical bitwise OR of two fixed-point vectors.
+   * @param[in]     pSrcA      points to input vector A
+   * @param[in]     pSrcB      points to input vector B
+   * @param[out]    pDst       points to output vector
+   * @param[in]     blockSize  number of samples in each vector
+   * @return        none
+   */
+  void arm_or_u8(
+    const uint8_t * pSrcA,
+    const uint8_t * pSrcB,
+          uint8_t * pDst,
+          uint32_t blockSize);
+
+  /**
+   * @brief         Compute the logical bitwise NOT of a fixed-point vector.
+   * @param[in]     pSrc       points to input vector 
+   * @param[out]    pDst       points to output vector
+   * @param[in]     blockSize  number of samples in each vector
+   * @return        none
+   */
+  void arm_not_u16(
+    const uint16_t * pSrc,
+          uint16_t * pDst,
+          uint32_t blockSize);
+
+  /**
+   * @brief         Compute the logical bitwise NOT of a fixed-point vector.
+   * @param[in]     pSrc       points to input vector 
+   * @param[out]    pDst       points to output vector
+   * @param[in]     blockSize  number of samples in each vector
+   * @return        none
+   */
+  void arm_not_u32(
+    const uint32_t * pSrc,
+          uint32_t * pDst,
+          uint32_t blockSize);
+
+  /**
+   * @brief         Compute the logical bitwise NOT of a fixed-point vector.
+   * @param[in]     pSrc       points to input vector 
+   * @param[out]    pDst       points to output vector
+   * @param[in]     blockSize  number of samples in each vector
+   * @return        none
+   */
+  void arm_not_u8(
+    const uint8_t * pSrc,
+          uint8_t * pDst,
+          uint32_t blockSize);
+
+/**
+   * @brief         Compute the logical bitwise XOR of two fixed-point vectors.
+   * @param[in]     pSrcA      points to input vector A
+   * @param[in]     pSrcB      points to input vector B
+   * @param[out]    pDst       points to output vector
+   * @param[in]     blockSize  number of samples in each vector
+   * @return        none
+   */
+  void arm_xor_u16(
+    const uint16_t * pSrcA,
+    const uint16_t * pSrcB,
+          uint16_t * pDst,
+          uint32_t blockSize);
+
+  /**
+   * @brief         Compute the logical bitwise XOR of two fixed-point vectors.
+   * @param[in]     pSrcA      points to input vector A
+   * @param[in]     pSrcB      points to input vector B
+   * @param[out]    pDst       points to output vector
+   * @param[in]     blockSize  number of samples in each vector
+   * @return        none
+   */
+  void arm_xor_u32(
+    const uint32_t * pSrcA,
+    const uint32_t * pSrcB,
+          uint32_t * pDst,
+          uint32_t blockSize);
+
+  /**
+   * @brief         Compute the logical bitwise XOR of two fixed-point vectors.
+   * @param[in]     pSrcA      points to input vector A
+   * @param[in]     pSrcB      points to input vector B
+   * @param[out]    pDst       points to output vector
+   * @param[in]     blockSize  number of samples in each vector
+   * @return        none
+   */
+  void arm_xor_u8(
+    const uint8_t * pSrcA,
+    const uint8_t * pSrcB,
+          uint8_t * pDst,
+    uint32_t blockSize);
+
+  /**
+  @brief         Elementwise floating-point clipping
+  @param[in]     pSrc          points to input values
+  @param[out]    pDst          points to output clipped values
+  @param[in]     low           lower bound
+  @param[in]     high          higher bound
+  @param[in]     numSamples    number of samples to clip
+  @return        none
+ */
+
+void arm_clip_f32(const float32_t * pSrc, 
+  float32_t * pDst, 
+  float32_t low, 
+  float32_t high, 
+  uint32_t numSamples);
+
+  /**
+  @brief         Elementwise fixed-point clipping
+  @param[in]     pSrc          points to input values
+  @param[out]    pDst          points to output clipped values
+  @param[in]     low           lower bound
+  @param[in]     high          higher bound
+  @param[in]     numSamples    number of samples to clip
+  @return        none
+ */
+
+void arm_clip_q31(const q31_t * pSrc, 
+  q31_t * pDst, 
+  q31_t low, 
+  q31_t high, 
+  uint32_t numSamples);
+
+  /**
+  @brief         Elementwise fixed-point clipping
+  @param[in]     pSrc          points to input values
+  @param[out]    pDst          points to output clipped values
+  @param[in]     low           lower bound
+  @param[in]     high          higher bound
+  @param[in]     numSamples    number of samples to clip
+  @return        none
+ */
+
+void arm_clip_q15(const q15_t * pSrc, 
+  q15_t * pDst, 
+  q15_t low, 
+  q15_t high, 
+  uint32_t numSamples);
+
+  /**
+  @brief         Elementwise fixed-point clipping
+  @param[in]     pSrc          points to input values
+  @param[out]    pDst          points to output clipped values
+  @param[in]     low           lower bound
+  @param[in]     high          higher bound
+  @param[in]     numSamples    number of samples to clip
+  @return        none
+ */
+
+void arm_clip_q7(const q7_t * pSrc, 
+  q7_t * pDst, 
+  q7_t low, 
+  q7_t high, 
+  uint32_t numSamples);
+
+
+#ifdef   __cplusplus
+}
+#endif
+
+#endif /* ifndef _BASIC_MATH_FUNCTIONS_H_ */

bsp/cmsis/include/dsp/basic_math_functions_f16.h

@@ -0,0 +1,168 @@
+/******************************************************************************
+ * @file     basic_math_functions_f16.h
+ * @brief    Public header file for CMSIS DSP Library
+ * @version  V1.10.0
+ * @date     08 July 2021
+ * Target Processor: Cortex-M and Cortex-A cores
+ ******************************************************************************/
+/*
+ * Copyright (c) 2010-2020 Arm Limited or its affiliates. All rights reserved.
+ *
+ * SPDX-License-Identifier: Apache-2.0
+ *
+ * Licensed under the Apache License, Version 2.0 (the License); you may
+ * not use this file except in compliance with the License.
+ * You may obtain a copy of the License at
+ *
+ * www.apache.org/licenses/LICENSE-2.0
+ *
+ * Unless required by applicable law or agreed to in writing, software
+ * distributed under the License is distributed on an AS IS BASIS, WITHOUT
+ * WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ * See the License for the specific language governing permissions and
+ * limitations under the License.
+ */
+
+ 
+#ifndef _BASIC_MATH_FUNCTIONS_F16_H_
+#define _BASIC_MATH_FUNCTIONS_F16_H_
+
+#ifdef   __cplusplus
+extern "C"
+{
+#endif
+
+#include "arm_math_types_f16.h"
+#include "arm_math_memory.h"
+
+#include "dsp/none.h"
+#include "dsp/utils.h"
+
+
+#if defined(ARM_FLOAT16_SUPPORTED)
+
+
+  /**
+   * @brief Floating-point vector addition.
+   * @param[in]  pSrcA      points to the first input vector
+   * @param[in]  pSrcB      points to the second input vector
+   * @param[out] pDst       points to the output vector
+   * @param[in]  blockSize  number of samples in each vector
+   */
+  void arm_add_f16(
+  const float16_t * pSrcA,
+  const float16_t * pSrcB,
+        float16_t * pDst,
+        uint32_t blockSize);
+
+  /**
+   * @brief Floating-point vector subtraction.
+   * @param[in]  pSrcA      points to the first input vector
+   * @param[in]  pSrcB      points to the second input vector
+   * @param[out] pDst       points to the output vector
+   * @param[in]  blockSize  number of samples in each vector
+   */
+  void arm_sub_f16(
+  const float16_t * pSrcA,
+  const float16_t * pSrcB,
+        float16_t * pDst,
+        uint32_t blockSize);
+
+    /**
+   * @brief Multiplies a floating-point vector by a scalar.
+   * @param[in]  pSrc       points to the input vector
+   * @param[in]  scale      scale factor to be applied
+   * @param[out] pDst       points to the output vector
+   * @param[in]  blockSize  number of samples in the vector
+   */
+  void arm_scale_f16(
+  const float16_t * pSrc,
+        float16_t scale,
+        float16_t * pDst,
+        uint32_t blockSize);
+
+    /**
+   * @brief Floating-point vector absolute value.
+   * @param[in]  pSrc       points to the input buffer
+   * @param[out] pDst       points to the output buffer
+   * @param[in]  blockSize  number of samples in each vector
+   */
+  void arm_abs_f16(
+  const float16_t * pSrc,
+        float16_t * pDst,
+        uint32_t blockSize);
+
+
+  /**
+   * @brief  Adds a constant offset to a floating-point vector.
+   * @param[in]  pSrc       points to the input vector
+   * @param[in]  offset     is the offset to be added
+   * @param[out] pDst       points to the output vector
+   * @param[in]  blockSize  number of samples in the vector
+   */
+  void arm_offset_f16(
+  const float16_t * pSrc,
+        float16_t offset,
+        float16_t * pDst,
+        uint32_t blockSize);
+
+  /**
+   * @brief Dot product of floating-point vectors.
+   * @param[in]  pSrcA      points to the first input vector
+   * @param[in]  pSrcB      points to the second input vector
+   * @param[in]  blockSize  number of samples in each vector
+   * @param[out] result     output result returned here
+   */
+  void arm_dot_prod_f16(
+  const float16_t * pSrcA,
+  const float16_t * pSrcB,
+        uint32_t blockSize,
+        float16_t * result);
+
+  /**
+   * @brief Floating-point vector multiplication.
+   * @param[in]  pSrcA      points to the first input vector
+   * @param[in]  pSrcB      points to the second input vector
+   * @param[out] pDst       points to the output vector
+   * @param[in]  blockSize  number of samples in each vector
+   */
+  void arm_mult_f16(
+  const float16_t * pSrcA,
+  const float16_t * pSrcB,
+        float16_t * pDst,
+        uint32_t blockSize);
+
+  /**
+   * @brief  Negates the elements of a floating-point vector.
+   * @param[in]  pSrc       points to the input vector
+   * @param[out] pDst       points to the output vector
+   * @param[in]  blockSize  number of samples in the vector
+   */
+  void arm_negate_f16(
+  const float16_t * pSrc,
+        float16_t * pDst,
+        uint32_t blockSize);
+
+  /**
+  @brief         Elementwise floating-point clipping
+  @param[in]     pSrc          points to input values
+  @param[out]    pDst          points to output clipped values
+  @param[in]     low           lower bound
+  @param[in]     high          higher bound
+  @param[in]     numSamples    number of samples to clip
+  @return        none
+ */
+
+void arm_clip_f16(const float16_t * pSrc, 
+  float16_t * pDst, 
+  float16_t low, 
+  float16_t high, 
+  uint32_t numSamples);
+
+#endif /* defined(ARM_FLOAT16_SUPPORTED)*/
+
+#ifdef   __cplusplus
+}
+#endif
+
+#endif /* ifndef _BASIC_MATH_FUNCTIONS_F16_H_ */

bsp/cmsis/include/dsp/bayes_functions.h

@@ -0,0 +1,89 @@
+/******************************************************************************
+ * @file     bayes_functions.h
+ * @brief    Public header file for CMSIS DSP Library
+ * @version  V1.10.0
+ * @date     08 July 2021
+ * Target Processor: Cortex-M and Cortex-A cores
+ ******************************************************************************/
+/*
+ * Copyright (c) 2010-2020 Arm Limited or its affiliates. All rights reserved.
+ *
+ * SPDX-License-Identifier: Apache-2.0
+ *
+ * Licensed under the Apache License, Version 2.0 (the License); you may
+ * not use this file except in compliance with the License.
+ * You may obtain a copy of the License at
+ *
+ * www.apache.org/licenses/LICENSE-2.0
+ *
+ * Unless required by applicable law or agreed to in writing, software
+ * distributed under the License is distributed on an AS IS BASIS, WITHOUT
+ * WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ * See the License for the specific language governing permissions and
+ * limitations under the License.
+ */
+
+ 
+#ifndef _BAYES_FUNCTIONS_H_
+#define _BAYES_FUNCTIONS_H_
+
+#include "arm_math_types.h"
+#include "arm_math_memory.h"
+
+#include "dsp/none.h"
+#include "dsp/utils.h"
+
+#include "dsp/statistics_functions.h"
+
+/**
+ * @defgroup groupBayes Bayesian estimators
+ *
+ * Implement the naive gaussian Bayes estimator.
+ * The training must be done from scikit-learn.
+ *
+ * The parameters can be easily
+ * generated from the scikit-learn object. Some examples are given in
+ * DSP/Testing/PatternGeneration/Bayes.py
+ */
+
+#ifdef   __cplusplus
+extern "C"
+{
+#endif
+
+/**
+ * @brief Instance structure for Naive Gaussian Bayesian estimator.
+ */
+typedef struct
+{
+  uint32_t vectorDimension;  /**< Dimension of vector space */
+  uint32_t numberOfClasses;  /**< Number of different classes  */
+  const float32_t *theta;          /**< Mean values for the Gaussians */
+  const float32_t *sigma;          /**< Variances for the Gaussians */
+  const float32_t *classPriors;    /**< Class prior probabilities */
+  float32_t epsilon;         /**< Additive value to variances */
+} arm_gaussian_naive_bayes_instance_f32;
+
+/**
+ * @brief Naive Gaussian Bayesian Estimator
+ *
+ * @param[in]  S                        points to a naive bayes instance structure
+ * @param[in]  in                       points to the elements of the input vector.
+ * @param[out] *pOutputProbabilities    points to a buffer of length numberOfClasses containing estimated probabilities
+ * @param[out] *pBufferB                points to a temporary buffer of length numberOfClasses
+ * @return The predicted class
+ *
+ */
+
+
+uint32_t arm_gaussian_naive_bayes_predict_f32(const arm_gaussian_naive_bayes_instance_f32 *S, 
+   const float32_t * in, 
+   float32_t *pOutputProbabilities,
+   float32_t *pBufferB);
+
+
+#ifdef   __cplusplus
+}
+#endif
+
+#endif /* ifndef _BAYES_FUNCTIONS_H_ */

bsp/cmsis/include/dsp/bayes_functions_f16.h

@@ -0,0 +1,80 @@
+/******************************************************************************
+ * @file     bayes_functions_f16.h
+ * @brief    Public header file for CMSIS DSP Library
+ * @version  V1.10.0
+ * @date     08 July 2021
+ * Target Processor: Cortex-M and Cortex-A cores
+ ******************************************************************************/
+/*
+ * Copyright (c) 2010-2020 Arm Limited or its affiliates. All rights reserved.
+ *
+ * SPDX-License-Identifier: Apache-2.0
+ *
+ * Licensed under the Apache License, Version 2.0 (the License); you may
+ * not use this file except in compliance with the License.
+ * You may obtain a copy of the License at
+ *
+ * www.apache.org/licenses/LICENSE-2.0
+ *
+ * Unless required by applicable law or agreed to in writing, software
+ * distributed under the License is distributed on an AS IS BASIS, WITHOUT
+ * WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ * See the License for the specific language governing permissions and
+ * limitations under the License.
+ */
+
+ 
+#ifndef _BAYES_FUNCTIONS_F16_H_
+#define _BAYES_FUNCTIONS_F16_H_
+
+#include "arm_math_types_f16.h"
+#include "arm_math_memory.h"
+
+#include "dsp/none.h"
+#include "dsp/utils.h"
+
+#include "dsp/statistics_functions_f16.h"
+
+#ifdef   __cplusplus
+extern "C"
+{
+#endif
+
+#if defined(ARM_FLOAT16_SUPPORTED)
+
+/**
+ * @brief Instance structure for Naive Gaussian Bayesian estimator.
+ */
+typedef struct
+{
+  uint32_t vectorDimension;  /**< Dimension of vector space */
+  uint32_t numberOfClasses;  /**< Number of different classes  */
+  const float16_t *theta;          /**< Mean values for the Gaussians */
+  const float16_t *sigma;          /**< Variances for the Gaussians */
+  const float16_t *classPriors;    /**< Class prior probabilities */
+  float16_t epsilon;         /**< Additive value to variances */
+} arm_gaussian_naive_bayes_instance_f16;
+
+/**
+ * @brief Naive Gaussian Bayesian Estimator
+ *
+ * @param[in]  S                        points to a naive bayes instance structure
+ * @param[in]  in                       points to the elements of the input vector.
+ * @param[out] *pOutputProbabilities    points to a buffer of length numberOfClasses containing estimated probabilities
+ * @param[out] *pBufferB                points to a temporary buffer of length numberOfClasses
+ * @return The predicted class
+ *
+ */
+
+
+uint32_t arm_gaussian_naive_bayes_predict_f16(const arm_gaussian_naive_bayes_instance_f16 *S, 
+   const float16_t * in, 
+   float16_t *pOutputProbabilities,
+   float16_t *pBufferB);
+
+#endif /*defined(ARM_FLOAT16_SUPPORTED)*/
+#ifdef   __cplusplus
+}
+#endif
+
+#endif /* ifndef _BAYES_FUNCTIONS_F16_H_ */

bsp/cmsis/include/dsp/complex_math_functions.h

@@ -0,0 +1,345 @@
+/******************************************************************************
+ * @file     complex_math_functions.h
+ * @brief    Public header file for CMSIS DSP Library
+ * @version  V1.10.0
+ * @date     08 July 2021
+ * Target Processor: Cortex-M and Cortex-A cores
+ ******************************************************************************/
+/*
+ * Copyright (c) 2010-2020 Arm Limited or its affiliates. All rights reserved.
+ *
+ * SPDX-License-Identifier: Apache-2.0
+ *
+ * Licensed under the Apache License, Version 2.0 (the License); you may
+ * not use this file except in compliance with the License.
+ * You may obtain a copy of the License at
+ *
+ * www.apache.org/licenses/LICENSE-2.0
+ *
+ * Unless required by applicable law or agreed to in writing, software
+ * distributed under the License is distributed on an AS IS BASIS, WITHOUT
+ * WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ * See the License for the specific language governing permissions and
+ * limitations under the License.
+ */
+
+ 
+#ifndef _COMPLEX_MATH_FUNCTIONS_H_
+#define _COMPLEX_MATH_FUNCTIONS_H_
+
+#include "arm_math_types.h"
+#include "arm_math_memory.h"
+
+#include "dsp/none.h"
+#include "dsp/utils.h"
+#include "dsp/fast_math_functions.h"
+
+#ifdef   __cplusplus
+extern "C"
+{
+#endif
+
+/**
+ * @defgroup groupCmplxMath Complex Math Functions
+ * This set of functions operates on complex data vectors.
+ * The data in the complex arrays is stored in an interleaved fashion
+ * (real, imag, real, imag, ...).
+ * In the API functions, the number of samples in a complex array refers
+ * to the number of complex values; the array contains twice this number of
+ * real values.
+ */
+
+ /**
+   * @brief  Floating-point complex conjugate.
+   * @param[in]  pSrc        points to the input vector
+   * @param[out] pDst        points to the output vector
+   * @param[in]  numSamples  number of complex samples in each vector
+   */
+  void arm_cmplx_conj_f32(
+  const float32_t * pSrc,
+        float32_t * pDst,
+        uint32_t numSamples);
+
+  /**
+   * @brief  Q31 complex conjugate.
+   * @param[in]  pSrc        points to the input vector
+   * @param[out] pDst        points to the output vector
+   * @param[in]  numSamples  number of complex samples in each vector
+   */
+  void arm_cmplx_conj_q31(
+  const q31_t * pSrc,
+        q31_t * pDst,
+        uint32_t numSamples);
+
+
+  /**
+   * @brief  Q15 complex conjugate.
+   * @param[in]  pSrc        points to the input vector
+   * @param[out] pDst        points to the output vector
+   * @param[in]  numSamples  number of complex samples in each vector
+   */
+  void arm_cmplx_conj_q15(
+  const q15_t * pSrc,
+        q15_t * pDst,
+        uint32_t numSamples);
+
+
+  /**
+   * @brief  Floating-point complex magnitude squared
+   * @param[in]  pSrc        points to the complex input vector
+   * @param[out] pDst        points to the real output vector
+   * @param[in]  numSamples  number of complex samples in the input vector
+   */
+  void arm_cmplx_mag_squared_f32(
+  const float32_t * pSrc,
+        float32_t * pDst,
+        uint32_t numSamples);
+
+
+  /**
+   * @brief  Floating-point complex magnitude squared
+   * @param[in]  pSrc        points to the complex input vector
+   * @param[out] pDst        points to the real output vector
+   * @param[in]  numSamples  number of complex samples in the input vector
+   */
+  void arm_cmplx_mag_squared_f64(
+  const float64_t * pSrc,
+        float64_t * pDst,
+        uint32_t numSamples);
+
+
+  /**
+   * @brief  Q31 complex magnitude squared
+   * @param[in]  pSrc        points to the complex input vector
+   * @param[out] pDst        points to the real output vector
+   * @param[in]  numSamples  number of complex samples in the input vector
+   */
+  void arm_cmplx_mag_squared_q31(
+  const q31_t * pSrc,
+        q31_t * pDst,
+        uint32_t numSamples);
+
+
+  /**
+   * @brief  Q15 complex magnitude squared
+   * @param[in]  pSrc        points to the complex input vector
+   * @param[out] pDst        points to the real output vector
+   * @param[in]  numSamples  number of complex samples in the input vector
+   */
+  void arm_cmplx_mag_squared_q15(
+  const q15_t * pSrc,
+        q15_t * pDst,
+        uint32_t numSamples);
+
+
+/**
+   * @brief  Floating-point complex magnitude
+   * @param[in]  pSrc        points to the complex input vector
+   * @param[out] pDst        points to the real output vector
+   * @param[in]  numSamples  number of complex samples in the input vector
+   */
+  void arm_cmplx_mag_f32(
+  const float32_t * pSrc,
+        float32_t * pDst,
+        uint32_t numSamples);
+
+
+/**
+   * @brief  Floating-point complex magnitude
+   * @param[in]  pSrc        points to the complex input vector
+   * @param[out] pDst        points to the real output vector
+   * @param[in]  numSamples  number of complex samples in the input vector
+   */
+  void arm_cmplx_mag_f64(
+  const float64_t * pSrc,
+        float64_t * pDst,
+        uint32_t numSamples);
+
+
+  /**
+   * @brief  Q31 complex magnitude
+   * @param[in]  pSrc        points to the complex input vector
+   * @param[out] pDst        points to the real output vector
+   * @param[in]  numSamples  number of complex samples in the input vector
+   */
+  void arm_cmplx_mag_q31(
+  const q31_t * pSrc,
+        q31_t * pDst,
+        uint32_t numSamples);
+
+
+  /**
+   * @brief  Q15 complex magnitude
+   * @param[in]  pSrc        points to the complex input vector
+   * @param[out] pDst        points to the real output vector
+   * @param[in]  numSamples  number of complex samples in the input vector
+   */
+  void arm_cmplx_mag_q15(
+  const q15_t * pSrc,
+        q15_t * pDst,
+        uint32_t numSamples);
+
+  /**
+   * @brief  Q15 complex magnitude
+   * @param[in]  pSrc        points to the complex input vector
+   * @param[out] pDst        points to the real output vector
+   * @param[in]  numSamples  number of complex samples in the input vector
+   */
+  void arm_cmplx_mag_fast_q15(
+  const q15_t * pSrc,
+        q15_t * pDst,
+        uint32_t numSamples);
+
+
+  /**
+   * @brief  Q15 complex dot product
+   * @param[in]  pSrcA       points to the first input vector
+   * @param[in]  pSrcB       points to the second input vector
+   * @param[in]  numSamples  number of complex samples in each vector
+   * @param[out] realResult  real part of the result returned here
+   * @param[out] imagResult  imaginary part of the result returned here
+   */
+  void arm_cmplx_dot_prod_q15(
+  const q15_t * pSrcA,
+  const q15_t * pSrcB,
+        uint32_t numSamples,
+        q31_t * realResult,
+        q31_t * imagResult);
+
+
+  /**
+   * @brief  Q31 complex dot product
+   * @param[in]  pSrcA       points to the first input vector
+   * @param[in]  pSrcB       points to the second input vector
+   * @param[in]  numSamples  number of complex samples in each vector
+   * @param[out] realResult  real part of the result returned here
+   * @param[out] imagResult  imaginary part of the result returned here
+   */
+  void arm_cmplx_dot_prod_q31(
+  const q31_t * pSrcA,
+  const q31_t * pSrcB,
+        uint32_t numSamples,
+        q63_t * realResult,
+        q63_t * imagResult);
+
+
+  /**
+   * @brief  Floating-point complex dot product
+   * @param[in]  pSrcA       points to the first input vector
+   * @param[in]  pSrcB       points to the second input vector
+   * @param[in]  numSamples  number of complex samples in each vector
+   * @param[out] realResult  real part of the result returned here
+   * @param[out] imagResult  imaginary part of the result returned here
+   */
+  void arm_cmplx_dot_prod_f32(
+  const float32_t * pSrcA,
+  const float32_t * pSrcB,
+        uint32_t numSamples,
+        float32_t * realResult,
+        float32_t * imagResult);
+
+
+  /**
+   * @brief  Q15 complex-by-real multiplication
+   * @param[in]  pSrcCmplx   points to the complex input vector
+   * @param[in]  pSrcReal    points to the real input vector
+   * @param[out] pCmplxDst   points to the complex output vector
+   * @param[in]  numSamples  number of samples in each vector
+   */
+  void arm_cmplx_mult_real_q15(
+  const q15_t * pSrcCmplx,
+  const q15_t * pSrcReal,
+        q15_t * pCmplxDst,
+        uint32_t numSamples);
+
+
+  /**
+   * @brief  Q31 complex-by-real multiplication
+   * @param[in]  pSrcCmplx   points to the complex input vector
+   * @param[in]  pSrcReal    points to the real input vector
+   * @param[out] pCmplxDst   points to the complex output vector
+   * @param[in]  numSamples  number of samples in each vector
+   */
+  void arm_cmplx_mult_real_q31(
+  const q31_t * pSrcCmplx,
+  const q31_t * pSrcReal,
+        q31_t * pCmplxDst,
+        uint32_t numSamples);
+
+
+  /**
+   * @brief  Floating-point complex-by-real multiplication
+   * @param[in]  pSrcCmplx   points to the complex input vector
+   * @param[in]  pSrcReal    points to the real input vector
+   * @param[out] pCmplxDst   points to the complex output vector
+   * @param[in]  numSamples  number of samples in each vector
+   */
+  void arm_cmplx_mult_real_f32(
+  const float32_t * pSrcCmplx,
+  const float32_t * pSrcReal,
+        float32_t * pCmplxDst,
+        uint32_t numSamples);
+
+  /**
+   * @brief  Q15 complex-by-complex multiplication
+   * @param[in]  pSrcA       points to the first input vector
+   * @param[in]  pSrcB       points to the second input vector
+   * @param[out] pDst        points to the output vector
+   * @param[in]  numSamples  number of complex samples in each vector
+   */
+  void arm_cmplx_mult_cmplx_q15(
+  const q15_t * pSrcA,
+  const q15_t * pSrcB,
+        q15_t * pDst,
+        uint32_t numSamples);
+
+
+  /**
+   * @brief  Q31 complex-by-complex multiplication
+   * @param[in]  pSrcA       points to the first input vector
+   * @param[in]  pSrcB       points to the second input vector
+   * @param[out] pDst        points to the output vector
+   * @param[in]  numSamples  number of complex samples in each vector
+   */
+  void arm_cmplx_mult_cmplx_q31(
+  const q31_t * pSrcA,
+  const q31_t * pSrcB,
+        q31_t * pDst,
+        uint32_t numSamples);
+
+
+  /**
+   * @brief  Floating-point complex-by-complex multiplication
+   * @param[in]  pSrcA       points to the first input vector
+   * @param[in]  pSrcB       points to the second input vector
+   * @param[out] pDst        points to the output vector
+   * @param[in]  numSamples  number of complex samples in each vector
+   */
+  void arm_cmplx_mult_cmplx_f32(
+  const float32_t * pSrcA,
+  const float32_t * pSrcB,
+        float32_t * pDst,
+        uint32_t numSamples);
+
+
+
+/**
+ * @brief  Floating-point complex-by-complex multiplication
+ * @param[in]  pSrcA       points to the first input vector
+ * @param[in]  pSrcB       points to the second input vector
+ * @param[out] pDst        points to the output vector
+ * @param[in]  numSamples  number of complex samples in each vector
+ */
+void arm_cmplx_mult_cmplx_f64(
+const float64_t * pSrcA,
+const float64_t * pSrcB,
+	  float64_t * pDst,
+	  uint32_t numSamples);
+
+
+
+#ifdef   __cplusplus
+}
+#endif
+
+#endif /* ifndef _COMPLEX_MATH_FUNCTIONS_H_ */

bsp/cmsis/include/dsp/complex_math_functions_f16.h

@@ -0,0 +1,123 @@
+/******************************************************************************
+ * @file     complex_math_functions_f16.h
+ * @brief    Public header file for CMSIS DSP Library
+ * @version  V1.10.0
+ * @date     08 July 2021
+ * Target Processor: Cortex-M and Cortex-A cores
+ ******************************************************************************/
+/*
+ * Copyright (c) 2010-2020 Arm Limited or its affiliates. All rights reserved.
+ *
+ * SPDX-License-Identifier: Apache-2.0
+ *
+ * Licensed under the Apache License, Version 2.0 (the License); you may
+ * not use this file except in compliance with the License.
+ * You may obtain a copy of the License at
+ *
+ * www.apache.org/licenses/LICENSE-2.0
+ *
+ * Unless required by applicable law or agreed to in writing, software
+ * distributed under the License is distributed on an AS IS BASIS, WITHOUT
+ * WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ * See the License for the specific language governing permissions and
+ * limitations under the License.
+ */
+
+ 
+#ifndef _COMPLEX_MATH_FUNCTIONS_F16_H_
+#define _COMPLEX_MATH_FUNCTIONS_F16_H_
+
+#include "arm_math_types_f16.h"
+#include "arm_math_memory.h"
+
+#include "dsp/none.h"
+#include "dsp/utils.h"
+#include "dsp/fast_math_functions_f16.h"
+
+#ifdef   __cplusplus
+extern "C"
+{
+#endif
+
+#if defined(ARM_FLOAT16_SUPPORTED)
+
+ /**
+   * @brief  Floating-point complex conjugate.
+   * @param[in]  pSrc        points to the input vector
+   * @param[out] pDst        points to the output vector
+   * @param[in]  numSamples  number of complex samples in each vector
+   */
+  void arm_cmplx_conj_f16(
+  const float16_t * pSrc,
+        float16_t * pDst,
+        uint32_t numSamples);
+
+ /**
+   * @brief  Floating-point complex magnitude squared
+   * @param[in]  pSrc        points to the complex input vector
+   * @param[out] pDst        points to the real output vector
+   * @param[in]  numSamples  number of complex samples in the input vector
+   */
+  void arm_cmplx_mag_squared_f16(
+  const float16_t * pSrc,
+        float16_t * pDst,
+        uint32_t numSamples);
+
+  /**
+   * @brief  Floating-point complex magnitude
+   * @param[in]  pSrc        points to the complex input vector
+   * @param[out] pDst        points to the real output vector
+   * @param[in]  numSamples  number of complex samples in the input vector
+   */
+  void arm_cmplx_mag_f16(
+  const float16_t * pSrc,
+        float16_t * pDst,
+        uint32_t numSamples);
+
+  /**
+   * @brief  Floating-point complex dot product
+   * @param[in]  pSrcA       points to the first input vector
+   * @param[in]  pSrcB       points to the second input vector
+   * @param[in]  numSamples  number of complex samples in each vector
+   * @param[out] realResult  real part of the result returned here
+   * @param[out] imagResult  imaginary part of the result returned here
+   */
+  void arm_cmplx_dot_prod_f16(
+  const float16_t * pSrcA,
+  const float16_t * pSrcB,
+        uint32_t numSamples,
+        float16_t * realResult,
+        float16_t * imagResult);
+
+   /**
+   * @brief  Floating-point complex-by-real multiplication
+   * @param[in]  pSrcCmplx   points to the complex input vector
+   * @param[in]  pSrcReal    points to the real input vector
+   * @param[out] pCmplxDst   points to the complex output vector
+   * @param[in]  numSamples  number of samples in each vector
+   */
+  void arm_cmplx_mult_real_f16(
+  const float16_t * pSrcCmplx,
+  const float16_t * pSrcReal,
+        float16_t * pCmplxDst,
+        uint32_t numSamples);
+
+  /**
+   * @brief  Floating-point complex-by-complex multiplication
+   * @param[in]  pSrcA       points to the first input vector
+   * @param[in]  pSrcB       points to the second input vector
+   * @param[out] pDst        points to the output vector
+   * @param[in]  numSamples  number of complex samples in each vector
+   */
+  void arm_cmplx_mult_cmplx_f16(
+  const float16_t * pSrcA,
+  const float16_t * pSrcB,
+        float16_t * pDst,
+        uint32_t numSamples);
+
+#endif /*defined(ARM_FLOAT16_SUPPORTED)*/
+#ifdef   __cplusplus
+}
+#endif
+
+#endif /* ifndef _COMPLEX_MATH_FUNCTIONS_F16_H_ */

bsp/cmsis/include/dsp/controller_functions.h

@@ -0,0 +1,791 @@
+/******************************************************************************
+ * @file     controller_functions.h
+ * @brief    Public header file for CMSIS DSP Library
+ * @version  V1.10.0
+ * @date     08 July 2021
+ * Target Processor: Cortex-M and Cortex-A cores
+ ******************************************************************************/
+/*
+ * Copyright (c) 2010-2020 Arm Limited or its affiliates. All rights reserved.
+ *
+ * SPDX-License-Identifier: Apache-2.0
+ *
+ * Licensed under the Apache License, Version 2.0 (the License); you may
+ * not use this file except in compliance with the License.
+ * You may obtain a copy of the License at
+ *
+ * www.apache.org/licenses/LICENSE-2.0
+ *
+ * Unless required by applicable law or agreed to in writing, software
+ * distributed under the License is distributed on an AS IS BASIS, WITHOUT
+ * WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ * See the License for the specific language governing permissions and
+ * limitations under the License.
+ */
+
+ 
+#ifndef _CONTROLLER_FUNCTIONS_H_
+#define _CONTROLLER_FUNCTIONS_H_
+
+#include "arm_math_types.h"
+#include "arm_math_memory.h"
+
+#include "dsp/none.h"
+#include "dsp/utils.h"
+
+#ifdef   __cplusplus
+extern "C"
+{
+#endif
+
+  /**
+   * @brief Macros required for SINE and COSINE Controller functions
+   */
+
+#define CONTROLLER_Q31_SHIFT  (32 - 9)
+  /* 1.31(q31) Fixed value of 2/360 */
+  /* -1 to +1 is divided into 360 values so total spacing is (2/360) */
+#define INPUT_SPACING         0xB60B61
+  
+/**
+ * @defgroup groupController Controller Functions
+ */
+
+
+ /**
+   * @ingroup groupController
+   */
+
+  /**
+   * @addtogroup SinCos
+   * @{
+   */
+
+/**
+   * @brief  Floating-point sin_cos function.
+   * @param[in]  theta   input value in degrees
+   * @param[out] pSinVal  points to the processed sine output.
+   * @param[out] pCosVal  points to the processed cos output.
+   */
+  void arm_sin_cos_f32(
+        float32_t theta,
+        float32_t * pSinVal,
+        float32_t * pCosVal);
+
+
+  /**
+   * @brief  Q31 sin_cos function.
+   * @param[in]  theta    scaled input value in degrees
+   * @param[out] pSinVal  points to the processed sine output.
+   * @param[out] pCosVal  points to the processed cosine output.
+   */
+  void arm_sin_cos_q31(
+        q31_t theta,
+        q31_t * pSinVal,
+        q31_t * pCosVal);
+
+  /**
+   * @} end of SinCos group
+   */
+
+ /**
+   * @ingroup groupController
+   */
+
+/**
+   * @defgroup PID PID Motor Control
+   *
+   * A Proportional Integral Derivative (PID) controller is a generic feedback control
+   * loop mechanism widely used in industrial control systems.
+   * A PID controller is the most commonly used type of feedback controller.
+   *
+   * This set of functions implements (PID) controllers
+   * for Q15, Q31, and floating-point data types.  The functions operate on a single sample
+   * of data and each call to the function returns a single processed value.
+   * <code>S</code> points to an instance of the PID control data structure.  <code>in</code>
+   * is the input sample value. The functions return the output value.
+   *
+   * \par Algorithm:
+   * <pre>
+   *    y[n] = y[n-1] + A0 * x[n] + A1 * x[n-1] + A2 * x[n-2]
+   *    A0 = Kp + Ki + Kd
+   *    A1 = (-Kp ) - (2 * Kd )
+   *    A2 = Kd
+   * </pre>
+   *
+   * \par
+   * where \c Kp is proportional constant, \c Ki is Integral constant and \c Kd is Derivative constant
+   *
+   * \par
+   * \image html PID.gif "Proportional Integral Derivative Controller"
+   *
+   * \par
+   * The PID controller calculates an "error" value as the difference between
+   * the measured output and the reference input.
+   * The controller attempts to minimize the error by adjusting the process control inputs.
+   * The proportional value determines the reaction to the current error,
+   * the integral value determines the reaction based on the sum of recent errors,
+   * and the derivative value determines the reaction based on the rate at which the error has been changing.
+   *
+   * \par Instance Structure
+   * The Gains A0, A1, A2 and state variables for a PID controller are stored together in an instance data structure.
+   * A separate instance structure must be defined for each PID Controller.
+   * There are separate instance structure declarations for each of the 3 supported data types.
+   *
+   * \par Reset Functions
+   * There is also an associated reset function for each data type which clears the state array.
+   *
+   * \par Initialization Functions
+   * There is also an associated initialization function for each data type.
+   * The initialization function performs the following operations:
+   * - Initializes the Gains A0, A1, A2 from Kp,Ki, Kd gains.
+   * - Zeros out the values in the state buffer.
+   *
+   * \par
+   * Instance structure cannot be placed into a const data section and it is recommended to use the initialization function.
+   *
+   * \par Fixed-Point Behavior
+   * Care must be taken when using the fixed-point versions of the PID Controller functions.
+   * In particular, the overflow and saturation behavior of the accumulator used in each function must be considered.
+   * Refer to the function specific documentation below for usage guidelines.
+   */
+
+
+  /**
+   * @brief Instance structure for the Q15 PID Control.
+   */
+  typedef struct
+  {
+          q15_t A0;           /**< The derived gain, A0 = Kp + Ki + Kd . */
+#if !defined (ARM_MATH_DSP)
+          q15_t A1;           /**< The derived gain A1 = -Kp - 2Kd */
+          q15_t A2;           /**< The derived gain A1 = Kd. */
+#else
+          q31_t A1;           /**< The derived gain A1 = -Kp - 2Kd | Kd.*/
+#endif
+          q15_t state[3];     /**< The state array of length 3. */
+          q15_t Kp;           /**< The proportional gain. */
+          q15_t Ki;           /**< The integral gain. */
+          q15_t Kd;           /**< The derivative gain. */
+  } arm_pid_instance_q15;
+
+  /**
+   * @brief Instance structure for the Q31 PID Control.
+   */
+  typedef struct
+  {
+          q31_t A0;            /**< The derived gain, A0 = Kp + Ki + Kd . */
+          q31_t A1;            /**< The derived gain, A1 = -Kp - 2Kd. */
+          q31_t A2;            /**< The derived gain, A2 = Kd . */
+          q31_t state[3];      /**< The state array of length 3. */
+          q31_t Kp;            /**< The proportional gain. */
+          q31_t Ki;            /**< The integral gain. */
+          q31_t Kd;            /**< The derivative gain. */
+  } arm_pid_instance_q31;
+
+  /**
+   * @brief Instance structure for the floating-point PID Control.
+   */
+  typedef struct
+  {
+          float32_t A0;          /**< The derived gain, A0 = Kp + Ki + Kd . */
+          float32_t A1;          /**< The derived gain, A1 = -Kp - 2Kd. */
+          float32_t A2;          /**< The derived gain, A2 = Kd . */
+          float32_t state[3];    /**< The state array of length 3. */
+          float32_t Kp;          /**< The proportional gain. */
+          float32_t Ki;          /**< The integral gain. */
+          float32_t Kd;          /**< The derivative gain. */
+  } arm_pid_instance_f32;
+
+
+
+  /**
+   * @brief  Initialization function for the floating-point PID Control.
+   * @param[in,out] S               points to an instance of the PID structure.
+   * @param[in]     resetStateFlag  flag to reset the state. 0 = no change in state 1 = reset the state.
+   */
+  void arm_pid_init_f32(
+        arm_pid_instance_f32 * S,
+        int32_t resetStateFlag);
+
+
+  /**
+   * @brief  Reset function for the floating-point PID Control.
+   * @param[in,out] S  is an instance of the floating-point PID Control structure
+   */
+  void arm_pid_reset_f32(
+        arm_pid_instance_f32 * S);
+
+
+  /**
+   * @brief  Initialization function for the Q31 PID Control.
+   * @param[in,out] S               points to an instance of the Q15 PID structure.
+   * @param[in]     resetStateFlag  flag to reset the state. 0 = no change in state 1 = reset the state.
+   */
+  void arm_pid_init_q31(
+        arm_pid_instance_q31 * S,
+        int32_t resetStateFlag);
+
+
+  /**
+   * @brief  Reset function for the Q31 PID Control.
+   * @param[in,out] S   points to an instance of the Q31 PID Control structure
+   */
+
+  void arm_pid_reset_q31(
+        arm_pid_instance_q31 * S);
+
+
+  /**
+   * @brief  Initialization function for the Q15 PID Control.
+   * @param[in,out] S               points to an instance of the Q15 PID structure.
+   * @param[in]     resetStateFlag  flag to reset the state. 0 = no change in state 1 = reset the state.
+   */
+  void arm_pid_init_q15(
+        arm_pid_instance_q15 * S,
+        int32_t resetStateFlag);
+
+
+  /**
+   * @brief  Reset function for the Q15 PID Control.
+   * @param[in,out] S  points to an instance of the q15 PID Control structure
+   */
+  void arm_pid_reset_q15(
+        arm_pid_instance_q15 * S);
+
+
+
+  /**
+   * @addtogroup PID
+   * @{
+   */
+
+  /**
+   * @brief         Process function for the floating-point PID Control.
+   * @param[in,out] S   is an instance of the floating-point PID Control structure
+   * @param[in]     in  input sample to process
+   * @return        processed output sample.
+   */
+  __STATIC_FORCEINLINE float32_t arm_pid_f32(
+  arm_pid_instance_f32 * S,
+  float32_t in)
+  {
+    float32_t out;
+
+    /* y[n] = y[n-1] + A0 * x[n] + A1 * x[n-1] + A2 * x[n-2]  */
+    out = (S->A0 * in) +
+      (S->A1 * S->state[0]) + (S->A2 * S->state[1]) + (S->state[2]);
+
+    /* Update state */
+    S->state[1] = S->state[0];
+    S->state[0] = in;
+    S->state[2] = out;
+
+    /* return to application */
+    return (out);
+
+  }
+
+/**
+  @brief         Process function for the Q31 PID Control.
+  @param[in,out] S  points to an instance of the Q31 PID Control structure
+  @param[in]     in  input sample to process
+  @return        processed output sample.
+
+  \par Scaling and Overflow Behavior
+         The function is implemented using an internal 64-bit accumulator.
+         The accumulator has a 2.62 format and maintains full precision of the intermediate multiplication results but provides only a single guard bit.
+         Thus, if the accumulator result overflows it wraps around rather than clip.
+         In order to avoid overflows completely the input signal must be scaled down by 2 bits as there are four additions.
+         After all multiply-accumulates are performed, the 2.62 accumulator is truncated to 1.32 format and then saturated to 1.31 format.
+ */
+__STATIC_FORCEINLINE q31_t arm_pid_q31(
+  arm_pid_instance_q31 * S,
+  q31_t in)
+  {
+    q63_t acc;
+    q31_t out;
+
+    /* acc = A0 * x[n]  */
+    acc = (q63_t) S->A0 * in;
+
+    /* acc += A1 * x[n-1] */
+    acc += (q63_t) S->A1 * S->state[0];
+
+    /* acc += A2 * x[n-2]  */
+    acc += (q63_t) S->A2 * S->state[1];
+
+    /* convert output to 1.31 format to add y[n-1] */
+    out = (q31_t) (acc >> 31U);
+
+    /* out += y[n-1] */
+    out += S->state[2];
+
+    /* Update state */
+    S->state[1] = S->state[0];
+    S->state[0] = in;
+    S->state[2] = out;
+
+    /* return to application */
+    return (out);
+  }
+
+
+/**
+  @brief         Process function for the Q15 PID Control.
+  @param[in,out] S   points to an instance of the Q15 PID Control structure
+  @param[in]     in  input sample to process
+  @return        processed output sample.
+
+  \par Scaling and Overflow Behavior
+         The function is implemented using a 64-bit internal accumulator.
+         Both Gains and state variables are represented in 1.15 format and multiplications yield a 2.30 result.
+         The 2.30 intermediate results are accumulated in a 64-bit accumulator in 34.30 format.
+         There is no risk of internal overflow with this approach and the full precision of intermediate multiplications is preserved.
+         After all additions have been performed, the accumulator is truncated to 34.15 format by discarding low 15 bits.
+         Lastly, the accumulator is saturated to yield a result in 1.15 format.
+ */
+__STATIC_FORCEINLINE q15_t arm_pid_q15(
+  arm_pid_instance_q15 * S,
+  q15_t in)
+  {
+    q63_t acc;
+    q15_t out;
+
+#if defined (ARM_MATH_DSP)
+    /* Implementation of PID controller */
+
+    /* acc = A0 * x[n]  */
+    acc = (q31_t) __SMUAD((uint32_t)S->A0, (uint32_t)in);
+
+    /* acc += A1 * x[n-1] + A2 * x[n-2]  */
+    acc = (q63_t)__SMLALD((uint32_t)S->A1, (uint32_t)read_q15x2 (S->state), (uint64_t)acc);
+#else
+    /* acc = A0 * x[n]  */
+    acc = ((q31_t) S->A0) * in;
+
+    /* acc += A1 * x[n-1] + A2 * x[n-2]  */
+    acc += (q31_t) S->A1 * S->state[0];
+    acc += (q31_t) S->A2 * S->state[1];
+#endif
+
+    /* acc += y[n-1] */
+    acc += (q31_t) S->state[2] << 15;
+
+    /* saturate the output */
+    out = (q15_t) (__SSAT((q31_t)(acc >> 15), 16));
+
+    /* Update state */
+    S->state[1] = S->state[0];
+    S->state[0] = in;
+    S->state[2] = out;
+
+    /* return to application */
+    return (out);
+  }
+
+  /**
+   * @} end of PID group
+   */
+
+  /**
+   * @ingroup groupController
+   */
+
+  /**
+   * @defgroup park Vector Park Transform
+   *
+   * Forward Park transform converts the input two-coordinate vector to flux and torque components.
+   * The Park transform can be used to realize the transformation of the <code>Ialpha</code> and the <code>Ibeta</code> currents
+   * from the stationary to the moving reference frame and control the spatial relationship between
+   * the stator vector current and rotor flux vector.
+   * If we consider the d axis aligned with the rotor flux, the diagram below shows the
+   * current vector and the relationship from the two reference frames:
+   * \image html park.gif "Stator current space vector and its component in (a,b) and in the d,q rotating reference frame"
+   *
+   * The function operates on a single sample of data and each call to the function returns the processed output.
+   * The library provides separate functions for Q31 and floating-point data types.
+   * \par Algorithm
+   * \image html parkFormula.gif
+   * where <code>Ialpha</code> and <code>Ibeta</code> are the stator vector components,
+   * <code>pId</code> and <code>pIq</code> are rotor vector components and <code>cosVal</code> and <code>sinVal</code> are the
+   * cosine and sine values of theta (rotor flux position).
+   * \par Fixed-Point Behavior
+   * Care must be taken when using the Q31 version of the Park transform.
+   * In particular, the overflow and saturation behavior of the accumulator used must be considered.
+   * Refer to the function specific documentation below for usage guidelines.
+   */
+
+  /**
+   * @addtogroup park
+   * @{
+   */
+
+  /**
+   * @brief Floating-point Park transform
+   * @param[in]  Ialpha  input two-phase vector coordinate alpha
+   * @param[in]  Ibeta   input two-phase vector coordinate beta
+   * @param[out] pId     points to output   rotor reference frame d
+   * @param[out] pIq     points to output   rotor reference frame q
+   * @param[in]  sinVal  sine value of rotation angle theta
+   * @param[in]  cosVal  cosine value of rotation angle theta
+   * @return     none
+   *
+   * The function implements the forward Park transform.
+   *
+   */
+  __STATIC_FORCEINLINE void arm_park_f32(
+  float32_t Ialpha,
+  float32_t Ibeta,
+  float32_t * pId,
+  float32_t * pIq,
+  float32_t sinVal,
+  float32_t cosVal)
+  {
+    /* Calculate pId using the equation, pId = Ialpha * cosVal + Ibeta * sinVal */
+    *pId = Ialpha * cosVal + Ibeta * sinVal;
+
+    /* Calculate pIq using the equation, pIq = - Ialpha * sinVal + Ibeta * cosVal */
+    *pIq = -Ialpha * sinVal + Ibeta * cosVal;
+  }
+
+
+/**
+  @brief  Park transform for Q31 version
+  @param[in]  Ialpha  input two-phase vector coordinate alpha
+  @param[in]  Ibeta   input two-phase vector coordinate beta
+  @param[out] pId     points to output rotor reference frame d
+  @param[out] pIq     points to output rotor reference frame q
+  @param[in]  sinVal  sine value of rotation angle theta
+  @param[in]  cosVal  cosine value of rotation angle theta
+  @return     none
+
+  \par Scaling and Overflow Behavior
+         The function is implemented using an internal 32-bit accumulator.
+         The accumulator maintains 1.31 format by truncating lower 31 bits of the intermediate multiplication in 2.62 format.
+         There is saturation on the addition and subtraction, hence there is no risk of overflow.
+ */
+__STATIC_FORCEINLINE void arm_park_q31(
+  q31_t Ialpha,
+  q31_t Ibeta,
+  q31_t * pId,
+  q31_t * pIq,
+  q31_t sinVal,
+  q31_t cosVal)
+  {
+    q31_t product1, product2;                    /* Temporary variables used to store intermediate results */
+    q31_t product3, product4;                    /* Temporary variables used to store intermediate results */
+
+    /* Intermediate product is calculated by (Ialpha * cosVal) */
+    product1 = (q31_t) (((q63_t) (Ialpha) * (cosVal)) >> 31);
+
+    /* Intermediate product is calculated by (Ibeta * sinVal) */
+    product2 = (q31_t) (((q63_t) (Ibeta) * (sinVal)) >> 31);
+
+
+    /* Intermediate product is calculated by (Ialpha * sinVal) */
+    product3 = (q31_t) (((q63_t) (Ialpha) * (sinVal)) >> 31);
+
+    /* Intermediate product is calculated by (Ibeta * cosVal) */
+    product4 = (q31_t) (((q63_t) (Ibeta) * (cosVal)) >> 31);
+
+    /* Calculate pId by adding the two intermediate products 1 and 2 */
+    *pId = __QADD(product1, product2);
+
+    /* Calculate pIq by subtracting the two intermediate products 3 from 4 */
+    *pIq = __QSUB(product4, product3);
+  }
+
+  /**
+   * @} end of park group
+   */
+
+
+  /**
+   * @ingroup groupController
+   */
+
+  /**
+   * @defgroup inv_park Vector Inverse Park transform
+   * Inverse Park transform converts the input flux and torque components to two-coordinate vector.
+   *
+   * The function operates on a single sample of data and each call to the function returns the processed output.
+   * The library provides separate functions for Q31 and floating-point data types.
+   * \par Algorithm
+   * \image html parkInvFormula.gif
+   * where <code>pIalpha</code> and <code>pIbeta</code> are the stator vector components,
+   * <code>Id</code> and <code>Iq</code> are rotor vector components and <code>cosVal</code> and <code>sinVal</code> are the
+   * cosine and sine values of theta (rotor flux position).
+   * \par Fixed-Point Behavior
+   * Care must be taken when using the Q31 version of the Park transform.
+   * In particular, the overflow and saturation behavior of the accumulator used must be considered.
+   * Refer to the function specific documentation below for usage guidelines.
+   */
+
+  /**
+   * @addtogroup inv_park
+   * @{
+   */
+
+   /**
+   * @brief  Floating-point Inverse Park transform
+   * @param[in]  Id       input coordinate of rotor reference frame d
+   * @param[in]  Iq       input coordinate of rotor reference frame q
+   * @param[out] pIalpha  points to output two-phase orthogonal vector axis alpha
+   * @param[out] pIbeta   points to output two-phase orthogonal vector axis beta
+   * @param[in]  sinVal   sine value of rotation angle theta
+   * @param[in]  cosVal   cosine value of rotation angle theta
+   * @return     none
+   */
+  __STATIC_FORCEINLINE void arm_inv_park_f32(
+  float32_t Id,
+  float32_t Iq,
+  float32_t * pIalpha,
+  float32_t * pIbeta,
+  float32_t sinVal,
+  float32_t cosVal)
+  {
+    /* Calculate pIalpha using the equation, pIalpha = Id * cosVal - Iq * sinVal */
+    *pIalpha = Id * cosVal - Iq * sinVal;
+
+    /* Calculate pIbeta using the equation, pIbeta = Id * sinVal + Iq * cosVal */
+    *pIbeta = Id * sinVal + Iq * cosVal;
+  }
+
+
+/**
+  @brief  Inverse Park transform for   Q31 version
+  @param[in]  Id       input coordinate of rotor reference frame d
+  @param[in]  Iq       input coordinate of rotor reference frame q
+  @param[out] pIalpha  points to output two-phase orthogonal vector axis alpha
+  @param[out] pIbeta   points to output two-phase orthogonal vector axis beta
+  @param[in]  sinVal   sine value of rotation angle theta
+  @param[in]  cosVal   cosine value of rotation angle theta
+  @return     none
+
+  @par Scaling and Overflow Behavior
+         The function is implemented using an internal 32-bit accumulator.
+         The accumulator maintains 1.31 format by truncating lower 31 bits of the intermediate multiplication in 2.62 format.
+         There is saturation on the addition, hence there is no risk of overflow.
+ */
+__STATIC_FORCEINLINE void arm_inv_park_q31(
+  q31_t Id,
+  q31_t Iq,
+  q31_t * pIalpha,
+  q31_t * pIbeta,
+  q31_t sinVal,
+  q31_t cosVal)
+  {
+    q31_t product1, product2;                    /* Temporary variables used to store intermediate results */
+    q31_t product3, product4;                    /* Temporary variables used to store intermediate results */
+
+    /* Intermediate product is calculated by (Id * cosVal) */
+    product1 = (q31_t) (((q63_t) (Id) * (cosVal)) >> 31);
+
+    /* Intermediate product is calculated by (Iq * sinVal) */
+    product2 = (q31_t) (((q63_t) (Iq) * (sinVal)) >> 31);
+
+
+    /* Intermediate product is calculated by (Id * sinVal) */
+    product3 = (q31_t) (((q63_t) (Id) * (sinVal)) >> 31);
+
+    /* Intermediate product is calculated by (Iq * cosVal) */
+    product4 = (q31_t) (((q63_t) (Iq) * (cosVal)) >> 31);
+
+    /* Calculate pIalpha by using the two intermediate products 1 and 2 */
+    *pIalpha = __QSUB(product1, product2);
+
+    /* Calculate pIbeta by using the two intermediate products 3 and 4 */
+    *pIbeta = __QADD(product4, product3);
+  }
+
+  /**
+   * @} end of Inverse park group
+   */
+
+/**
+   * @ingroup groupController
+   */
+
+  /**
+   * @defgroup clarke Vector Clarke Transform
+   * Forward Clarke transform converts the instantaneous stator phases into a two-coordinate time invariant vector.
+   * Generally the Clarke transform uses three-phase currents <code>Ia, Ib and Ic</code> to calculate currents
+   * in the two-phase orthogonal stator axis <code>Ialpha</code> and <code>Ibeta</code>.
+   * When <code>Ialpha</code> is superposed with <code>Ia</code> as shown in the figure below
+   * \image html clarke.gif Stator current space vector and its components in (a,b).
+   * and <code>Ia + Ib + Ic = 0</code>, in this condition <code>Ialpha</code> and <code>Ibeta</code>
+   * can be calculated using only <code>Ia</code> and <code>Ib</code>.
+   *
+   * The function operates on a single sample of data and each call to the function returns the processed output.
+   * The library provides separate functions for Q31 and floating-point data types.
+   * \par Algorithm
+   * \image html clarkeFormula.gif
+   * where <code>Ia</code> and <code>Ib</code> are the instantaneous stator phases and
+   * <code>pIalpha</code> and <code>pIbeta</code> are the two coordinates of time invariant vector.
+   * \par Fixed-Point Behavior
+   * Care must be taken when using the Q31 version of the Clarke transform.
+   * In particular, the overflow and saturation behavior of the accumulator used must be considered.
+   * Refer to the function specific documentation below for usage guidelines.
+   */
+
+  /**
+   * @addtogroup clarke
+   * @{
+   */
+
+  /**
+   *
+   * @brief  Floating-point Clarke transform
+   * @param[in]  Ia       input three-phase coordinate <code>a</code>
+   * @param[in]  Ib       input three-phase coordinate <code>b</code>
+   * @param[out] pIalpha  points to output two-phase orthogonal vector axis alpha
+   * @param[out] pIbeta   points to output two-phase orthogonal vector axis beta
+   * @return        none
+   */
+  __STATIC_FORCEINLINE void arm_clarke_f32(
+  float32_t Ia,
+  float32_t Ib,
+  float32_t * pIalpha,
+  float32_t * pIbeta)
+  {
+    /* Calculate pIalpha using the equation, pIalpha = Ia */
+    *pIalpha = Ia;
+
+    /* Calculate pIbeta using the equation, pIbeta = (1/sqrt(3)) * Ia + (2/sqrt(3)) * Ib */
+    *pIbeta = (0.57735026919f * Ia + 1.15470053838f * Ib);
+  }
+
+
+/**
+  @brief  Clarke transform for Q31 version
+  @param[in]  Ia       input three-phase coordinate <code>a</code>
+  @param[in]  Ib       input three-phase coordinate <code>b</code>
+  @param[out] pIalpha  points to output two-phase orthogonal vector axis alpha
+  @param[out] pIbeta   points to output two-phase orthogonal vector axis beta
+  @return     none
+
+  \par Scaling and Overflow Behavior
+         The function is implemented using an internal 32-bit accumulator.
+         The accumulator maintains 1.31 format by truncating lower 31 bits of the intermediate multiplication in 2.62 format.
+         There is saturation on the addition, hence there is no risk of overflow.
+ */
+__STATIC_FORCEINLINE void arm_clarke_q31(
+  q31_t Ia,
+  q31_t Ib,
+  q31_t * pIalpha,
+  q31_t * pIbeta)
+  {
+    q31_t product1, product2;                    /* Temporary variables used to store intermediate results */
+
+    /* Calculating pIalpha from Ia by equation pIalpha = Ia */
+    *pIalpha = Ia;
+
+    /* Intermediate product is calculated by (1/(sqrt(3)) * Ia) */
+    product1 = (q31_t) (((q63_t) Ia * 0x24F34E8B) >> 30);
+
+    /* Intermediate product is calculated by (2/sqrt(3) * Ib) */
+    product2 = (q31_t) (((q63_t) Ib * 0x49E69D16) >> 30);
+
+    /* pIbeta is calculated by adding the intermediate products */
+    *pIbeta = __QADD(product1, product2);
+  }
+
+  /**
+   * @} end of clarke group
+   */
+
+
+  /**
+   * @ingroup groupController
+   */
+
+  /**
+   * @defgroup inv_clarke Vector Inverse Clarke Transform
+   * Inverse Clarke transform converts the two-coordinate time invariant vector into instantaneous stator phases.
+   *
+   * The function operates on a single sample of data and each call to the function returns the processed output.
+   * The library provides separate functions for Q31 and floating-point data types.
+   * \par Algorithm
+   * \image html clarkeInvFormula.gif
+   * where <code>pIa</code> and <code>pIb</code> are the instantaneous stator phases and
+   * <code>Ialpha</code> and <code>Ibeta</code> are the two coordinates of time invariant vector.
+   * \par Fixed-Point Behavior
+   * Care must be taken when using the Q31 version of the Clarke transform.
+   * In particular, the overflow and saturation behavior of the accumulator used must be considered.
+   * Refer to the function specific documentation below for usage guidelines.
+   */
+
+  /**
+   * @addtogroup inv_clarke
+   * @{
+   */
+
+   /**
+   * @brief  Floating-point Inverse Clarke transform
+   * @param[in]  Ialpha  input two-phase orthogonal vector axis alpha
+   * @param[in]  Ibeta   input two-phase orthogonal vector axis beta
+   * @param[out] pIa     points to output three-phase coordinate <code>a</code>
+   * @param[out] pIb     points to output three-phase coordinate <code>b</code>
+   * @return     none
+   */
+  __STATIC_FORCEINLINE void arm_inv_clarke_f32(
+  float32_t Ialpha,
+  float32_t Ibeta,
+  float32_t * pIa,
+  float32_t * pIb)
+  {
+    /* Calculating pIa from Ialpha by equation pIa = Ialpha */
+    *pIa = Ialpha;
+
+    /* Calculating pIb from Ialpha and Ibeta by equation pIb = -(1/2) * Ialpha + (sqrt(3)/2) * Ibeta */
+    *pIb = -0.5f * Ialpha + 0.8660254039f * Ibeta;
+  }
+
+
+/**
+  @brief  Inverse Clarke transform for Q31 version
+  @param[in]  Ialpha  input two-phase orthogonal vector axis alpha
+  @param[in]  Ibeta   input two-phase orthogonal vector axis beta
+  @param[out] pIa     points to output three-phase coordinate <code>a</code>
+  @param[out] pIb     points to output three-phase coordinate <code>b</code>
+  @return     none
+
+  \par Scaling and Overflow Behavior
+         The function is implemented using an internal 32-bit accumulator.
+         The accumulator maintains 1.31 format by truncating lower 31 bits of the intermediate multiplication in 2.62 format.
+         There is saturation on the subtraction, hence there is no risk of overflow.
+ */
+__STATIC_FORCEINLINE void arm_inv_clarke_q31(
+  q31_t Ialpha,
+  q31_t Ibeta,
+  q31_t * pIa,
+  q31_t * pIb)
+  {
+    q31_t product1, product2;                    /* Temporary variables used to store intermediate results */
+
+    /* Calculating pIa from Ialpha by equation pIa = Ialpha */
+    *pIa = Ialpha;
+
+    /* Intermediate product is calculated by (1/(2*sqrt(3)) * Ia) */
+    product1 = (q31_t) (((q63_t) (Ialpha) * (0x40000000)) >> 31);
+
+    /* Intermediate product is calculated by (1/sqrt(3) * pIb) */
+    product2 = (q31_t) (((q63_t) (Ibeta) * (0x6ED9EBA1)) >> 31);
+
+    /* pIb is calculated by subtracting the products */
+    *pIb = __QSUB(product2, product1);
+  }
+
+  /**
+   * @} end of inv_clarke group
+   */
+
+
+
+  
+#ifdef   __cplusplus
+}
+#endif
+
+#endif /* ifndef _CONTROLLER_FUNCTIONS_H_ */

bsp/cmsis/include/dsp/controller_functions_f16.h

@@ -0,0 +1,41 @@
+/******************************************************************************
+ * @file     controller_functions_f16.h
+ * @brief    Public header file for CMSIS DSP Library
+ * @version  V1.10.0
+ * @date     08 July 2021
+ * Target Processor: Cortex-M and Cortex-A cores
+ ******************************************************************************/
+/*
+ * Copyright (c) 2010-2020 Arm Limited or its affiliates. All rights reserved.
+ *
+ * SPDX-License-Identifier: Apache-2.0
+ *
+ * Licensed under the Apache License, Version 2.0 (the License); you may
+ * not use this file except in compliance with the License.
+ * You may obtain a copy of the License at
+ *
+ * www.apache.org/licenses/LICENSE-2.0
+ *
+ * Unless required by applicable law or agreed to in writing, software
+ * distributed under the License is distributed on an AS IS BASIS, WITHOUT
+ * WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ * See the License for the specific language governing permissions and
+ * limitations under the License.
+ */
+
+ 
+#ifndef _CONTROLLER_FUNCTIONS_F16_H_
+#define _CONTROLLER_FUNCTIONS_F16_H_
+
+#ifdef   __cplusplus
+extern "C"
+{
+#endif
+
+#if defined(ARM_FLOAT16_SUPPORTED)
+#endif /*defined(ARM_FLOAT16_SUPPORTED)*/
+#ifdef   __cplusplus
+}
+#endif
+
+#endif /* ifndef _CONTROLLER_FUNCTIONS_F16_H_ */

bsp/cmsis/include/dsp/distance_functions.h

@@ -0,0 +1,341 @@
+/******************************************************************************
+ * @file     distance_functions.h
+ * @brief    Public header file for CMSIS DSP Library
+ * @version  V1.10.0
+ * @date     08 July 2021
+ * Target Processor: Cortex-M and Cortex-A cores
+ ******************************************************************************/
+/*
+ * Copyright (c) 2010-2020 Arm Limited or its affiliates. All rights reserved.
+ *
+ * SPDX-License-Identifier: Apache-2.0
+ *
+ * Licensed under the Apache License, Version 2.0 (the License); you may
+ * not use this file except in compliance with the License.
+ * You may obtain a copy of the License at
+ *
+ * www.apache.org/licenses/LICENSE-2.0
+ *
+ * Unless required by applicable law or agreed to in writing, software
+ * distributed under the License is distributed on an AS IS BASIS, WITHOUT
+ * WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ * See the License for the specific language governing permissions and
+ * limitations under the License.
+ */
+
+ 
+#ifndef _DISTANCE_FUNCTIONS_H_
+#define _DISTANCE_FUNCTIONS_H_
+
+#include "arm_math_types.h"
+#include "arm_math_memory.h"
+
+#include "dsp/none.h"
+#include "dsp/utils.h"
+
+#include "dsp/statistics_functions.h"
+#include "dsp/basic_math_functions.h"
+#include "dsp/fast_math_functions.h"
+
+#ifdef   __cplusplus
+extern "C"
+{
+#endif
+
+
+/**
+ * @defgroup groupDistance Distance functions
+ *
+ * Distance functions for use with clustering algorithms.
+ * There are distance functions for float vectors and boolean vectors.
+ *
+ */
+
+/* 6.14 bug */
+#if defined (__ARMCC_VERSION) && (__ARMCC_VERSION >= 6100100) && (__ARMCC_VERSION < 6150001)
+ 
+__attribute__((weak)) float __powisf2(float a, int b);
+
+#endif 
+
+/**
+ * @brief        Euclidean distance between two vectors
+ * @param[in]    pA         First vector
+ * @param[in]    pB         Second vector
+ * @param[in]    blockSize  vector length
+ * @return distance
+ *
+ */
+
+float32_t arm_euclidean_distance_f32(const float32_t *pA,const float32_t *pB, uint32_t blockSize);
+
+/**
+ * @brief        Euclidean distance between two vectors
+ * @param[in]    pA         First vector
+ * @param[in]    pB         Second vector
+ * @param[in]    blockSize  vector length
+ * @return distance
+ *
+ */
+
+float64_t arm_euclidean_distance_f64(const float64_t *pA,const float64_t *pB, uint32_t blockSize);
+
+/**
+ * @brief        Bray-Curtis distance between two vectors
+ * @param[in]    pA         First vector
+ * @param[in]    pB         Second vector
+ * @param[in]    blockSize  vector length
+ * @return distance
+ *
+ */
+float32_t arm_braycurtis_distance_f32(const float32_t *pA,const float32_t *pB, uint32_t blockSize);
+
+/**
+ * @brief        Canberra distance between two vectors
+ *
+ * This function may divide by zero when samples pA[i] and pB[i] are both zero.
+ * The result of the computation will be correct. So the division per zero may be
+ * ignored.
+ *
+ * @param[in]    pA         First vector
+ * @param[in]    pB         Second vector
+ * @param[in]    blockSize  vector length
+ * @return distance
+ *
+ */
+float32_t arm_canberra_distance_f32(const float32_t *pA,const float32_t *pB, uint32_t blockSize);
+
+
+/**
+ * @brief        Chebyshev distance between two vectors
+ * @param[in]    pA         First vector
+ * @param[in]    pB         Second vector
+ * @param[in]    blockSize  vector length
+ * @return distance
+ *
+ */
+float32_t arm_chebyshev_distance_f32(const float32_t *pA,const float32_t *pB, uint32_t blockSize);
+
+
+/**
+ * @brief        Chebyshev distance between two vectors
+ * @param[in]    pA         First vector
+ * @param[in]    pB         Second vector
+ * @param[in]    blockSize  vector length
+ * @return distance
+ *
+ */
+float64_t arm_chebyshev_distance_f64(const float64_t *pA,const float64_t *pB, uint32_t blockSize);
+
+
+/**
+ * @brief        Cityblock (Manhattan) distance between two vectors
+ * @param[in]    pA         First vector
+ * @param[in]    pB         Second vector
+ * @param[in]    blockSize  vector length
+ * @return distance
+ *
+ */
+float32_t arm_cityblock_distance_f32(const float32_t *pA,const float32_t *pB, uint32_t blockSize);
+
+/**
+ * @brief        Cityblock (Manhattan) distance between two vectors
+ * @param[in]    pA         First vector
+ * @param[in]    pB         Second vector
+ * @param[in]    blockSize  vector length
+ * @return distance
+ *
+ */
+float64_t arm_cityblock_distance_f64(const float64_t *pA,const float64_t *pB, uint32_t blockSize);
+
+/**
+ * @brief        Correlation distance between two vectors
+ *
+ * The input vectors are modified in place !
+ *
+ * @param[in]    pA         First vector
+ * @param[in]    pB         Second vector
+ * @param[in]    blockSize  vector length
+ * @return distance
+ *
+ */
+float32_t arm_correlation_distance_f32(float32_t *pA,float32_t *pB, uint32_t blockSize);
+
+/**
+ * @brief        Cosine distance between two vectors
+ *
+ * @param[in]    pA         First vector
+ * @param[in]    pB         Second vector
+ * @param[in]    blockSize  vector length
+ * @return distance
+ *
+ */
+
+float32_t arm_cosine_distance_f32(const float32_t *pA,const float32_t *pB, uint32_t blockSize);
+
+/**
+ * @brief        Cosine distance between two vectors
+ *
+ * @param[in]    pA         First vector
+ * @param[in]    pB         Second vector
+ * @param[in]    blockSize  vector length
+ * @return distance
+ *
+ */
+
+float64_t arm_cosine_distance_f64(const float64_t *pA,const float64_t *pB, uint32_t blockSize);
+
+/**
+ * @brief        Jensen-Shannon distance between two vectors
+ *
+ * This function is assuming that elements of second vector are > 0
+ * and 0 only when the corresponding element of first vector is 0.
+ * Otherwise the result of the computation does not make sense
+ * and for speed reasons, the cases returning NaN or Infinity are not
+ * managed.
+ *
+ * When the function is computing x log (x / y) with x 0 and y 0,
+ * it will compute the right value (0) but a division per zero will occur
+ * and shoudl be ignored in client code.
+ *
+ * @param[in]    pA         First vector
+ * @param[in]    pB         Second vector
+ * @param[in]    blockSize  vector length
+ * @return distance
+ *
+ */
+
+float32_t arm_jensenshannon_distance_f32(const float32_t *pA,const float32_t *pB,uint32_t blockSize);
+
+/**
+ * @brief        Minkowski distance between two vectors
+ *
+ * @param[in]    pA         First vector
+ * @param[in]    pB         Second vector
+ * @param[in]    n          Norm order (>= 2)
+ * @param[in]    blockSize  vector length
+ * @return distance
+ *
+ */
+
+
+
+float32_t arm_minkowski_distance_f32(const float32_t *pA,const float32_t *pB, int32_t order, uint32_t blockSize);
+
+/**
+ * @brief        Dice distance between two vectors
+ *
+ * @param[in]    pA              First vector of packed booleans
+ * @param[in]    pB              Second vector of packed booleans
+ * @param[in]    order           Distance order
+ * @param[in]    blockSize       Number of samples
+ * @return distance
+ *
+ */
+
+
+float32_t arm_dice_distance(const uint32_t *pA, const uint32_t *pB, uint32_t numberOfBools);
+
+/**
+ * @brief        Hamming distance between two vectors
+ *
+ * @param[in]    pA              First vector of packed booleans
+ * @param[in]    pB              Second vector of packed booleans
+ * @param[in]    numberOfBools   Number of booleans
+ * @return distance
+ *
+ */
+
+float32_t arm_hamming_distance(const uint32_t *pA, const uint32_t *pB, uint32_t numberOfBools);
+
+/**
+ * @brief        Jaccard distance between two vectors
+ *
+ * @param[in]    pA              First vector of packed booleans
+ * @param[in]    pB              Second vector of packed booleans
+ * @param[in]    numberOfBools   Number of booleans
+ * @return distance
+ *
+ */
+
+float32_t arm_jaccard_distance(const uint32_t *pA, const uint32_t *pB, uint32_t numberOfBools);
+
+/**
+ * @brief        Kulsinski distance between two vectors
+ *
+ * @param[in]    pA              First vector of packed booleans
+ * @param[in]    pB              Second vector of packed booleans
+ * @param[in]    numberOfBools   Number of booleans
+ * @return distance
+ *
+ */
+
+float32_t arm_kulsinski_distance(const uint32_t *pA, const uint32_t *pB, uint32_t numberOfBools);
+
+/**
+ * @brief        Roger Stanimoto distance between two vectors
+ *
+ * @param[in]    pA              First vector of packed booleans
+ * @param[in]    pB              Second vector of packed booleans
+ * @param[in]    numberOfBools   Number of booleans
+ * @return distance
+ *
+ */
+
+float32_t arm_rogerstanimoto_distance(const uint32_t *pA, const uint32_t *pB, uint32_t numberOfBools);
+
+/**
+ * @brief        Russell-Rao distance between two vectors
+ *
+ * @param[in]    pA              First vector of packed booleans
+ * @param[in]    pB              Second vector of packed booleans
+ * @param[in]    numberOfBools   Number of booleans
+ * @return distance
+ *
+ */
+
+float32_t arm_russellrao_distance(const uint32_t *pA, const uint32_t *pB, uint32_t numberOfBools);
+
+/**
+ * @brief        Sokal-Michener distance between two vectors
+ *
+ * @param[in]    pA              First vector of packed booleans
+ * @param[in]    pB              Second vector of packed booleans
+ * @param[in]    numberOfBools   Number of booleans
+ * @return distance
+ *
+ */
+
+float32_t arm_sokalmichener_distance(const uint32_t *pA, const uint32_t *pB, uint32_t numberOfBools);
+
+/**
+ * @brief        Sokal-Sneath distance between two vectors
+ *
+ * @param[in]    pA              First vector of packed booleans
+ * @param[in]    pB              Second vector of packed booleans
+ * @param[in]    numberOfBools   Number of booleans
+ * @return distance
+ *
+ */
+
+float32_t arm_sokalsneath_distance(const uint32_t *pA, const uint32_t *pB, uint32_t numberOfBools);
+
+/**
+ * @brief        Yule distance between two vectors
+ *
+ * @param[in]    pA              First vector of packed booleans
+ * @param[in]    pB              Second vector of packed booleans
+ * @param[in]    numberOfBools   Number of booleans
+ * @return distance
+ *
+ */
+
+float32_t arm_yule_distance(const uint32_t *pA, const uint32_t *pB, uint32_t numberOfBools);
+
+
+
+#ifdef   __cplusplus
+}
+#endif
+
+#endif /* ifndef _DISTANCE_FUNCTIONS_H_ */

bsp/cmsis/include/dsp/distance_functions_f16.h

@@ -0,0 +1,180 @@
+/******************************************************************************
+ * @file     distance_functions_f16.h
+ * @brief    Public header file for CMSIS DSP Library
+ * @version  V1.10.0
+ * @date     08 July 2021
+ * Target Processor: Cortex-M and Cortex-A cores
+ ******************************************************************************/
+/*
+ * Copyright (c) 2010-2020 Arm Limited or its affiliates. All rights reserved.
+ *
+ * SPDX-License-Identifier: Apache-2.0
+ *
+ * Licensed under the Apache License, Version 2.0 (the License); you may
+ * not use this file except in compliance with the License.
+ * You may obtain a copy of the License at
+ *
+ * www.apache.org/licenses/LICENSE-2.0
+ *
+ * Unless required by applicable law or agreed to in writing, software
+ * distributed under the License is distributed on an AS IS BASIS, WITHOUT
+ * WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ * See the License for the specific language governing permissions and
+ * limitations under the License.
+ */
+
+ 
+#ifndef _DISTANCE_FUNCTIONS_F16_H_
+#define _DISTANCE_FUNCTIONS_F16_H_
+
+#include "arm_math_types_f16.h"
+#include "arm_math_memory.h"
+
+#include "dsp/none.h"
+#include "dsp/utils.h"
+
+/* 6.14 bug */
+#if defined (__ARMCC_VERSION) && (__ARMCC_VERSION >= 6100100) && (__ARMCC_VERSION < 6150001)
+/* Defined in minkowski_f32 */
+__attribute__((weak)) float __powisf2(float a, int b);
+#endif 
+
+#include "dsp/statistics_functions_f16.h"
+#include "dsp/basic_math_functions_f16.h"
+
+#include "dsp/fast_math_functions_f16.h"
+
+#ifdef   __cplusplus
+extern "C"
+{
+#endif
+
+#if defined(ARM_FLOAT16_SUPPORTED)
+
+/**
+ * @brief        Euclidean distance between two vectors
+ * @param[in]    pA         First vector
+ * @param[in]    pB         Second vector
+ * @param[in]    blockSize  vector length
+ * @return distance
+ *
+ */
+
+float16_t arm_euclidean_distance_f16(const float16_t *pA,const float16_t *pB, uint32_t blockSize);
+
+/**
+ * @brief        Bray-Curtis distance between two vectors
+ * @param[in]    pA         First vector
+ * @param[in]    pB         Second vector
+ * @param[in]    blockSize  vector length
+ * @return distance
+ *
+ */
+float16_t arm_braycurtis_distance_f16(const float16_t *pA,const float16_t *pB, uint32_t blockSize);
+
+/**
+ * @brief        Canberra distance between two vectors
+ *
+ * This function may divide by zero when samples pA[i] and pB[i] are both zero.
+ * The result of the computation will be correct. So the division per zero may be
+ * ignored.
+ *
+ * @param[in]    pA         First vector
+ * @param[in]    pB         Second vector
+ * @param[in]    blockSize  vector length
+ * @return distance
+ *
+ */
+float16_t arm_canberra_distance_f16(const float16_t *pA,const float16_t *pB, uint32_t blockSize);
+
+
+/**
+ * @brief        Chebyshev distance between two vectors
+ * @param[in]    pA         First vector
+ * @param[in]    pB         Second vector
+ * @param[in]    blockSize  vector length
+ * @return distance
+ *
+ */
+float16_t arm_chebyshev_distance_f16(const float16_t *pA,const float16_t *pB, uint32_t blockSize);
+
+
+/**
+ * @brief        Cityblock (Manhattan) distance between two vectors
+ * @param[in]    pA         First vector
+ * @param[in]    pB         Second vector
+ * @param[in]    blockSize  vector length
+ * @return distance
+ *
+ */
+float16_t arm_cityblock_distance_f16(const float16_t *pA,const float16_t *pB, uint32_t blockSize);
+
+/**
+ * @brief        Correlation distance between two vectors
+ *
+ * The input vectors are modified in place !
+ *
+ * @param[in]    pA         First vector
+ * @param[in]    pB         Second vector
+ * @param[in]    blockSize  vector length
+ * @return distance
+ *
+ */
+float16_t arm_correlation_distance_f16(float16_t *pA,float16_t *pB, uint32_t blockSize);
+
+/**
+ * @brief        Cosine distance between two vectors
+ *
+ * @param[in]    pA         First vector
+ * @param[in]    pB         Second vector
+ * @param[in]    blockSize  vector length
+ * @return distance
+ *
+ */
+
+float16_t arm_cosine_distance_f16(const float16_t *pA,const float16_t *pB, uint32_t blockSize);
+
+/**
+ * @brief        Jensen-Shannon distance between two vectors
+ *
+ * This function is assuming that elements of second vector are > 0
+ * and 0 only when the corresponding element of first vector is 0.
+ * Otherwise the result of the computation does not make sense
+ * and for speed reasons, the cases returning NaN or Infinity are not
+ * managed.
+ *
+ * When the function is computing x log (x / y) with x 0 and y 0,
+ * it will compute the right value (0) but a division per zero will occur
+ * and shoudl be ignored in client code.
+ *
+ * @param[in]    pA         First vector
+ * @param[in]    pB         Second vector
+ * @param[in]    blockSize  vector length
+ * @return distance
+ *
+ */
+
+float16_t arm_jensenshannon_distance_f16(const float16_t *pA,const float16_t *pB,uint32_t blockSize);
+
+/**
+ * @brief        Minkowski distance between two vectors
+ *
+ * @param[in]    pA         First vector
+ * @param[in]    pB         Second vector
+ * @param[in]    n          Norm order (>= 2)
+ * @param[in]    blockSize  vector length
+ * @return distance
+ *
+ */
+
+
+
+float16_t arm_minkowski_distance_f16(const float16_t *pA,const float16_t *pB, int32_t order, uint32_t blockSize);
+
+
+#endif /*defined(ARM_FLOAT16_SUPPORTED)*/
+#ifdef   __cplusplus
+}
+#endif
+
+#endif /* ifndef _DISTANCE_FUNCTIONS_F16_H_ */

bsp/cmsis/include/dsp/fast_math_functions.h

@@ -0,0 +1,389 @@
+/******************************************************************************
+ * @file     fast_math_functions.h
+ * @brief    Public header file for CMSIS DSP Library
+ * @version  V1.10.0
+ * @date     08 July 2021
+ * Target Processor: Cortex-M and Cortex-A cores
+ ******************************************************************************/
+/*
+ * Copyright (c) 2010-2020 Arm Limited or its affiliates. All rights reserved.
+ *
+ * SPDX-License-Identifier: Apache-2.0
+ *
+ * Licensed under the Apache License, Version 2.0 (the License); you may
+ * not use this file except in compliance with the License.
+ * You may obtain a copy of the License at
+ *
+ * www.apache.org/licenses/LICENSE-2.0
+ *
+ * Unless required by applicable law or agreed to in writing, software
+ * distributed under the License is distributed on an AS IS BASIS, WITHOUT
+ * WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ * See the License for the specific language governing permissions and
+ * limitations under the License.
+ */
+
+ 
+#ifndef _FAST_MATH_FUNCTIONS_H_
+#define _FAST_MATH_FUNCTIONS_H_
+
+#include "arm_math_types.h"
+#include "arm_math_memory.h"
+
+#include "dsp/none.h"
+#include "dsp/utils.h"
+
+#include "dsp/basic_math_functions.h"
+
+
+#ifdef   __cplusplus
+extern "C"
+{
+#endif
+
+  /**
+   * @brief Macros required for SINE and COSINE Fast math approximations
+   */
+
+#define FAST_MATH_TABLE_SIZE  512
+#define FAST_MATH_Q31_SHIFT   (32 - 10)
+#define FAST_MATH_Q15_SHIFT   (16 - 10)
+  
+#ifndef PI
+  #define PI               3.14159265358979f
+#endif
+
+
+/**
+ * @defgroup groupFastMath Fast Math Functions
+ * This set of functions provides a fast approximation to sine, cosine, and square root.
+ * As compared to most of the other functions in the CMSIS math library, the fast math functions
+ * operate on individual values and not arrays.
+ * There are separate functions for Q15, Q31, and floating-point data.
+ *
+ */
+
+  /**
+   * @ingroup groupFastMath
+   */
+
+
+/**
+  @addtogroup sin
+  @{
+ */
+
+/**
+   * @brief  Fast approximation to the trigonometric sine function for floating-point data.
+   * @param[in] x  input value in radians.
+   * @return  sin(x).
+   */
+  float32_t arm_sin_f32(
+  float32_t x);
+
+
+  /**
+   * @brief  Fast approximation to the trigonometric sine function for Q31 data.
+   * @param[in] x  Scaled input value in radians.
+   * @return  sin(x).
+   */
+  q31_t arm_sin_q31(
+  q31_t x);
+
+
+  /**
+   * @brief  Fast approximation to the trigonometric sine function for Q15 data.
+   * @param[in] x  Scaled input value in radians.
+   * @return  sin(x).
+   */
+  q15_t arm_sin_q15(
+  q15_t x);
+
+/**
+  @} end of sin group
+ */
+
+/**
+  @addtogroup cos
+  @{
+ */
+
+  /**
+   * @brief  Fast approximation to the trigonometric cosine function for floating-point data.
+   * @param[in] x  input value in radians.
+   * @return  cos(x).
+   */
+  float32_t arm_cos_f32(
+  float32_t x);
+
+
+  /**
+   * @brief Fast approximation to the trigonometric cosine function for Q31 data.
+   * @param[in] x  Scaled input value in radians.
+   * @return  cos(x).
+   */
+  q31_t arm_cos_q31(
+  q31_t x);
+
+
+  /**
+   * @brief  Fast approximation to the trigonometric cosine function for Q15 data.
+   * @param[in] x  Scaled input value in radians.
+   * @return  cos(x).
+   */
+  q15_t arm_cos_q15(
+  q15_t x);
+
+/**
+  @} end of cos group
+ */
+
+
+/**
+  @brief         Floating-point vector of log values.
+  @param[in]     pSrc       points to the input vector
+  @param[out]    pDst       points to the output vector
+  @param[in]     blockSize  number of samples in each vector
+  @return        none
+ */
+  void arm_vlog_f32(
+  const float32_t * pSrc,
+        float32_t * pDst,
+        uint32_t blockSize);
+
+
+
+/**
+  @brief         Floating-point vector of log values.
+  @param[in]     pSrc       points to the input vector
+  @param[out]    pDst       points to the output vector
+  @param[in]     blockSize  number of samples in each vector
+  @return        none
+ */
+  void arm_vlog_f64(
+  const float64_t * pSrc,
+		float64_t * pDst,
+		uint32_t blockSize);
+
+
+
+  /**
+   * @brief  q31 vector of log values.
+   * @param[in]     pSrc       points to the input vector in q31
+   * @param[out]    pDst       points to the output vector in q5.26
+   * @param[in]     blockSize  number of samples in each vector
+   * @return        none
+   */
+  void arm_vlog_q31(const q31_t * pSrc,
+        q31_t * pDst,
+        uint32_t blockSize);
+
+  /**
+   * @brief  q15 vector of log values.
+   * @param[in]     pSrc       points to the input vector in q15
+   * @param[out]    pDst       points to the output vector in q4.11
+   * @param[in]     blockSize  number of samples in each vector
+   * @return        none
+   */
+  void arm_vlog_q15(const q15_t * pSrc,
+        q15_t * pDst,
+        uint32_t blockSize);
+
+
+
+/**
+  @brief         Floating-point vector of exp values.
+  @param[in]     pSrc       points to the input vector
+  @param[out]    pDst       points to the output vector
+  @param[in]     blockSize  number of samples in each vector
+  @return        none
+ */
+  void arm_vexp_f32(
+  const float32_t * pSrc,
+        float32_t * pDst,
+        uint32_t blockSize);
+
+
+
+/**
+  @brief         Floating-point vector of exp values.
+  @param[in]     pSrc       points to the input vector
+  @param[out]    pDst       points to the output vector
+  @param[in]     blockSize  number of samples in each vector
+  @return        none
+ */
+  void arm_vexp_f64(
+  const float64_t * pSrc,
+		float64_t * pDst,
+		uint32_t blockSize);
+
+
+
+ /**
+   * @defgroup SQRT Square Root
+   *
+   * Computes the square root of a number.
+   * There are separate functions for Q15, Q31, and floating-point data types.
+   * The square root function is computed using the Newton-Raphson algorithm.
+   * This is an iterative algorithm of the form:
+   * <pre>
+   *      x1 = x0 - f(x0)/f'(x0)
+   * </pre>
+   * where <code>x1</code> is the current estimate,
+   * <code>x0</code> is the previous estimate, and
+   * <code>f'(x0)</code> is the derivative of <code>f()</code> evaluated at <code>x0</code>.
+   * For the square root function, the algorithm reduces to:
+   * <pre>
+   *     x0 = in/2                         [initial guess]
+   *     x1 = 1/2 * ( x0 + in / x0)        [each iteration]
+   * </pre>
+   */
+
+
+  /**
+   * @addtogroup SQRT
+   * @{
+   */
+
+/**
+  @brief         Floating-point square root function.
+  @param[in]     in    input value
+  @param[out]    pOut  square root of input value
+  @return        execution status
+                   - \ref ARM_MATH_SUCCESS        : input value is positive
+                   - \ref ARM_MATH_ARGUMENT_ERROR : input value is negative; *pOut is set to 0
+ */
+__STATIC_FORCEINLINE arm_status arm_sqrt_f32(
+  const float32_t in,
+  float32_t * pOut)
+  {
+    if (in >= 0.0f)
+    {
+#if defined ( __CC_ARM )
+  #if defined __TARGET_FPU_VFP
+      *pOut = __sqrtf(in);
+  #else
+      *pOut = sqrtf(in);
+  #endif
+
+#elif defined ( __ICCARM__ )
+  #if defined __ARMVFP__
+      __ASM("VSQRT.F32 %0,%1" : "=t"(*pOut) : "t"(in));
+  #else
+      *pOut = sqrtf(in);
+  #endif
+
+#else
+      *pOut = sqrtf(in);
+#endif
+
+      return (ARM_MATH_SUCCESS);
+    }
+    else
+    {
+      *pOut = 0.0f;
+      return (ARM_MATH_ARGUMENT_ERROR);
+    }
+  }
+
+
+/**
+  @brief         Q31 square root function.
+  @param[in]     in    input value.  The range of the input value is [0 +1) or 0x00000000 to 0x7FFFFFFF
+  @param[out]    pOut  points to square root of input value
+  @return        execution status
+                   - \ref ARM_MATH_SUCCESS        : input value is positive
+                   - \ref ARM_MATH_ARGUMENT_ERROR : input value is negative; *pOut is set to 0
+ */
+arm_status arm_sqrt_q31(
+  q31_t in,
+  q31_t * pOut);
+
+
+/**
+  @brief         Q15 square root function.
+  @param[in]     in    input value.  The range of the input value is [0 +1) or 0x0000 to 0x7FFF
+  @param[out]    pOut  points to square root of input value
+  @return        execution status
+                   - \ref ARM_MATH_SUCCESS        : input value is positive
+                   - \ref ARM_MATH_ARGUMENT_ERROR : input value is negative; *pOut is set to 0
+ */
+arm_status arm_sqrt_q15(
+  q15_t in,
+  q15_t * pOut);
+
+
+
+  /**
+   * @} end of SQRT group
+   */
+
+  /**
+  @brief         Fixed point division
+  @param[in]     numerator    Numerator
+  @param[in]     denominator  Denominator
+  @param[out]    quotient     Quotient value normalized between -1.0 and 1.0
+  @param[out]    shift        Shift left value to get the unnormalized quotient
+  @return        error status
+
+  When dividing by 0, an error ARM_MATH_NANINF is returned. And the quotient is forced
+  to the saturated negative or positive value.
+ */
+
+arm_status arm_divide_q15(q15_t numerator,
+  q15_t denominator,
+  q15_t *quotient,
+  int16_t *shift);
+
+  /**
+  @brief         Fixed point division
+  @param[in]     numerator    Numerator
+  @param[in]     denominator  Denominator
+  @param[out]    quotient     Quotient value normalized between -1.0 and 1.0
+  @param[out]    shift        Shift left value to get the unnormalized quotient
+  @return        error status
+
+  When dividing by 0, an error ARM_MATH_NANINF is returned. And the quotient is forced
+  to the saturated negative or positive value.
+ */
+
+arm_status arm_divide_q31(q31_t numerator,
+  q31_t denominator,
+  q31_t *quotient,
+  int16_t *shift);
+
+
+
+  /**
+     @brief  Arc tangent in radian of y/x using sign of x and y to determine right quadrant.
+     @param[in]   y  y coordinate
+     @param[in]   x  x coordinate
+     @param[out]  result  Result
+     @return  error status.
+   */
+  arm_status arm_atan2_f32(float32_t y,float32_t x,float32_t *result);
+
+
+  /**
+     @brief  Arc tangent in radian of y/x using sign of x and y to determine right quadrant.
+     @param[in]   y  y coordinate
+     @param[in]   x  x coordinate
+     @param[out]  result  Result in Q2.29
+     @return  error status.
+   */
+  arm_status arm_atan2_q31(q31_t y,q31_t x,q31_t *result);
+
+  /**
+     @brief  Arc tangent in radian of y/x using sign of x and y to determine right quadrant.
+     @param[in]   y  y coordinate
+     @param[in]   x  x coordinate
+     @param[out]  result  Result in Q2.13
+     @return  error status.
+   */
+  arm_status arm_atan2_q15(q15_t y,q15_t x,q15_t *result);
+
+#ifdef   __cplusplus
+}
+#endif
+
+#endif /* ifndef _FAST_MATH_FUNCTIONS_H_ */

bsp/cmsis/include/dsp/fast_math_functions_f16.h

@@ -0,0 +1,125 @@
+/******************************************************************************
+ * @file     fast_math_functions_f16.h
+ * @brief    Public header file for CMSIS DSP Library
+ * @version  V1.10.0
+ * @date     08 July 2021
+ * Target Processor: Cortex-M and Cortex-A cores
+ ******************************************************************************/
+/*
+ * Copyright (c) 2010-2020 Arm Limited or its affiliates. All rights reserved.
+ *
+ * SPDX-License-Identifier: Apache-2.0
+ *
+ * Licensed under the Apache License, Version 2.0 (the License); you may
+ * not use this file except in compliance with the License.
+ * You may obtain a copy of the License at
+ *
+ * www.apache.org/licenses/LICENSE-2.0
+ *
+ * Unless required by applicable law or agreed to in writing, software
+ * distributed under the License is distributed on an AS IS BASIS, WITHOUT
+ * WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ * See the License for the specific language governing permissions and
+ * limitations under the License.
+ */
+
+ 
+#ifndef _FAST_MATH_FUNCTIONS_F16_H_
+#define _FAST_MATH_FUNCTIONS_F16_H_
+
+#include "arm_math_types_f16.h"
+#include "arm_math_memory.h"
+
+#include "dsp/none.h"
+#include "dsp/utils.h"
+
+/* For sqrt_f32 */
+#include "dsp/fast_math_functions.h"
+
+#ifdef   __cplusplus
+extern "C"
+{
+#endif
+
+#if defined(ARM_FLOAT16_SUPPORTED)
+
+ /**
+   * @addtogroup SQRT
+   * @{
+   */
+
+/**
+  @brief         Floating-point square root function.
+  @param[in]     in    input value
+  @param[out]    pOut  square root of input value
+  @return        execution status
+                   - \ref ARM_MATH_SUCCESS        : input value is positive
+                   - \ref ARM_MATH_ARGUMENT_ERROR : input value is negative; *pOut is set to 0
+ */
+__STATIC_FORCEINLINE arm_status arm_sqrt_f16(
+  float16_t in,
+  float16_t * pOut)
+  {
+    float32_t r;
+    arm_status status;
+    status=arm_sqrt_f32((float32_t)in,&r);
+    *pOut=(float16_t)r;
+    return(status);
+  }
+
+
+/**
+  @} end of SQRT group
+ */
+  
+/**
+  @brief         Floating-point vector of log values.
+  @param[in]     pSrc       points to the input vector
+  @param[out]    pDst       points to the output vector
+  @param[in]     blockSize  number of samples in each vector
+  @return        none
+ */
+  void arm_vlog_f16(
+  const float16_t * pSrc,
+        float16_t * pDst,
+        uint32_t blockSize);
+
+/**
+  @brief         Floating-point vector of exp values.
+  @param[in]     pSrc       points to the input vector
+  @param[out]    pDst       points to the output vector
+  @param[in]     blockSize  number of samples in each vector
+  @return        none
+ */
+  void arm_vexp_f16(
+  const float16_t * pSrc,
+        float16_t * pDst,
+        uint32_t blockSize);
+
+  /**
+  @brief         Floating-point vector of inverse values.
+  @param[in]     pSrc       points to the input vector
+  @param[out]    pDst       points to the output vector
+  @param[in]     blockSize  number of samples in each vector
+  @return        none
+ */
+  void arm_vinverse_f16(
+  const float16_t * pSrc,
+        float16_t * pDst,
+        uint32_t blockSize);
+
+  /**
+     @brief  Arc tangent in radian of y/x using sign of x and y to determine right quadrant.
+     @param[in]   y  y coordinate
+     @param[in]   x  x coordinate
+     @param[out]  result  Result
+     @return  error status.
+   */
+  arm_status arm_atan2_f16(float16_t y,float16_t x,float16_t *result);
+
+#endif /*defined(ARM_FLOAT16_SUPPORTED)*/
+#ifdef   __cplusplus
+}
+#endif
+
+#endif /* ifndef _FAST_MATH_FUNCTIONS_F16_H_ */

bsp/cmsis/include/dsp/filtering_functions.h

@@ -0,0 +1,2529 @@
+/******************************************************************************
+ * @file     filtering_functions.h
+ * @brief    Public header file for CMSIS DSP Library
+ * @version  V1.10.0
+ * @date     08 July 2021
+ * Target Processor: Cortex-M and Cortex-A cores
+ ******************************************************************************/
+/*
+ * Copyright (c) 2010-2020 Arm Limited or its affiliates. All rights reserved.
+ *
+ * SPDX-License-Identifier: Apache-2.0
+ *
+ * Licensed under the Apache License, Version 2.0 (the License); you may
+ * not use this file except in compliance with the License.
+ * You may obtain a copy of the License at
+ *
+ * www.apache.org/licenses/LICENSE-2.0
+ *
+ * Unless required by applicable law or agreed to in writing, software
+ * distributed under the License is distributed on an AS IS BASIS, WITHOUT
+ * WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ * See the License for the specific language governing permissions and
+ * limitations under the License.
+ */
+
+ 
+#ifndef _FILTERING_FUNCTIONS_H_
+#define _FILTERING_FUNCTIONS_H_
+
+#include "arm_math_types.h"
+#include "arm_math_memory.h"
+
+#include "dsp/none.h"
+#include "dsp/utils.h"
+
+#include "dsp/support_functions.h"
+#include "dsp/fast_math_functions.h"
+
+#ifdef   __cplusplus
+extern "C"
+{
+#endif
+
+
+
+#define DELTA_Q31          ((q31_t)(0x100))
+#define DELTA_Q15          ((q15_t)0x5)
+
+/**
+ * @defgroup groupFilters Filtering Functions
+ */
+    
+  /**
+   * @brief Instance structure for the Q7 FIR filter.
+   */
+  typedef struct
+  {
+          uint16_t numTaps;        /**< number of filter coefficients in the filter. */
+          q7_t *pState;            /**< points to the state variable array. The array is of length numTaps+blockSize-1. */
+    const q7_t *pCoeffs;           /**< points to the coefficient array. The array is of length numTaps.*/
+  } arm_fir_instance_q7;
+
+  /**
+   * @brief Instance structure for the Q15 FIR filter.
+   */
+  typedef struct
+  {
+          uint16_t numTaps;         /**< number of filter coefficients in the filter. */
+          q15_t *pState;            /**< points to the state variable array. The array is of length numTaps+blockSize-1. */
+    const q15_t *pCoeffs;           /**< points to the coefficient array. The array is of length numTaps.*/
+  } arm_fir_instance_q15;
+
+  /**
+   * @brief Instance structure for the Q31 FIR filter.
+   */
+  typedef struct
+  {
+          uint16_t numTaps;         /**< number of filter coefficients in the filter. */
+          q31_t *pState;            /**< points to the state variable array. The array is of length numTaps+blockSize-1. */
+    const q31_t *pCoeffs;           /**< points to the coefficient array. The array is of length numTaps. */
+  } arm_fir_instance_q31;
+
+  /**
+   * @brief Instance structure for the floating-point FIR filter.
+   */
+  typedef struct
+  {
+          uint16_t numTaps;     /**< number of filter coefficients in the filter. */
+          float32_t *pState;    /**< points to the state variable array. The array is of length numTaps+blockSize-1. */
+    const float32_t *pCoeffs;   /**< points to the coefficient array. The array is of length numTaps. */
+  } arm_fir_instance_f32;
+
+  /**
+   * @brief Instance structure for the floating-point FIR filter.
+   */
+  typedef struct
+  {
+          uint16_t numTaps;     /**< number of filter coefficients in the filter. */
+          float64_t *pState;    /**< points to the state variable array. The array is of length numTaps+blockSize-1. */
+    const float64_t *pCoeffs;   /**< points to the coefficient array. The array is of length numTaps. */
+  } arm_fir_instance_f64;
+
+  /**
+   * @brief Processing function for the Q7 FIR filter.
+   * @param[in]  S          points to an instance of the Q7 FIR filter structure.
+   * @param[in]  pSrc       points to the block of input data.
+   * @param[out] pDst       points to the block of output data.
+   * @param[in]  blockSize  number of samples to process.
+   */
+  void arm_fir_q7(
+  const arm_fir_instance_q7 * S,
+  const q7_t * pSrc,
+        q7_t * pDst,
+        uint32_t blockSize);
+
+  /**
+   * @brief  Initialization function for the Q7 FIR filter.
+   * @param[in,out] S          points to an instance of the Q7 FIR structure.
+   * @param[in]     numTaps    Number of filter coefficients in the filter.
+   * @param[in]     pCoeffs    points to the filter coefficients.
+   * @param[in]     pState     points to the state buffer.
+   * @param[in]     blockSize  number of samples that are processed.
+   *
+   * For the MVE version, the coefficient length must be a multiple of 16.
+   * You can pad with zeros if you have less coefficients.
+   */
+  void arm_fir_init_q7(
+        arm_fir_instance_q7 * S,
+        uint16_t numTaps,
+  const q7_t * pCoeffs,
+        q7_t * pState,
+        uint32_t blockSize);
+
+  /**
+   * @brief Processing function for the Q15 FIR filter.
+   * @param[in]  S          points to an instance of the Q15 FIR structure.
+   * @param[in]  pSrc       points to the block of input data.
+   * @param[out] pDst       points to the block of output data.
+   * @param[in]  blockSize  number of samples to process.
+   */
+  void arm_fir_q15(
+  const arm_fir_instance_q15 * S,
+  const q15_t * pSrc,
+        q15_t * pDst,
+        uint32_t blockSize);
+
+  /**
+   * @brief Processing function for the fast Q15 FIR filter (fast version).
+   * @param[in]  S          points to an instance of the Q15 FIR filter structure.
+   * @param[in]  pSrc       points to the block of input data.
+   * @param[out] pDst       points to the block of output data.
+   * @param[in]  blockSize  number of samples to process.
+   */
+  void arm_fir_fast_q15(
+  const arm_fir_instance_q15 * S,
+  const q15_t * pSrc,
+        q15_t * pDst,
+        uint32_t blockSize);
+
+  /**
+   * @brief  Initialization function for the Q15 FIR filter.
+   * @param[in,out] S          points to an instance of the Q15 FIR filter structure.
+   * @param[in]     numTaps    Number of filter coefficients in the filter. Must be even and greater than or equal to 4.
+   * @param[in]     pCoeffs    points to the filter coefficients.
+   * @param[in]     pState     points to the state buffer.
+   * @param[in]     blockSize  number of samples that are processed at a time.
+   * @return     The function returns either
+   * <code>ARM_MATH_SUCCESS</code> if initialization was successful or
+   * <code>ARM_MATH_ARGUMENT_ERROR</code> if <code>numTaps</code> is not a supported value.
+   *
+   * For the MVE version, the coefficient length must be a multiple of 8.
+   * You can pad with zeros if you have less coefficients.
+   *
+   */
+  arm_status arm_fir_init_q15(
+        arm_fir_instance_q15 * S,
+        uint16_t numTaps,
+  const q15_t * pCoeffs,
+        q15_t * pState,
+        uint32_t blockSize);
+
+  /**
+   * @brief Processing function for the Q31 FIR filter.
+   * @param[in]  S          points to an instance of the Q31 FIR filter structure.
+   * @param[in]  pSrc       points to the block of input data.
+   * @param[out] pDst       points to the block of output data.
+   * @param[in]  blockSize  number of samples to process.
+   */
+  void arm_fir_q31(
+  const arm_fir_instance_q31 * S,
+  const q31_t * pSrc,
+        q31_t * pDst,
+        uint32_t blockSize);
+
+  /**
+   * @brief Processing function for the fast Q31 FIR filter (fast version).
+   * @param[in]  S          points to an instance of the Q31 FIR filter structure.
+   * @param[in]  pSrc       points to the block of input data.
+   * @param[out] pDst       points to the block of output data.
+   * @param[in]  blockSize  number of samples to process.
+   */
+  void arm_fir_fast_q31(
+  const arm_fir_instance_q31 * S,
+  const q31_t * pSrc,
+        q31_t * pDst,
+        uint32_t blockSize);
+
+  /**
+   * @brief  Initialization function for the Q31 FIR filter.
+   * @param[in,out] S          points to an instance of the Q31 FIR structure.
+   * @param[in]     numTaps    Number of filter coefficients in the filter.
+   * @param[in]     pCoeffs    points to the filter coefficients.
+   * @param[in]     pState     points to the state buffer.
+   * @param[in]     blockSize  number of samples that are processed at a time.
+   *
+   * For the MVE version, the coefficient length must be a multiple of 4.
+   * You can pad with zeros if you have less coefficients.
+   */
+  void arm_fir_init_q31(
+        arm_fir_instance_q31 * S,
+        uint16_t numTaps,
+  const q31_t * pCoeffs,
+        q31_t * pState,
+        uint32_t blockSize);
+
+  /**
+   * @brief Processing function for the floating-point FIR filter.
+   * @param[in]  S          points to an instance of the floating-point FIR structure.
+   * @param[in]  pSrc       points to the block of input data.
+   * @param[out] pDst       points to the block of output data.
+   * @param[in]  blockSize  number of samples to process.
+   */
+  void arm_fir_f32(
+  const arm_fir_instance_f32 * S,
+  const float32_t * pSrc,
+        float32_t * pDst,
+        uint32_t blockSize);
+
+  /**
+   * @brief Processing function for the floating-point FIR filter.
+   * @param[in]  S          points to an instance of the floating-point FIR structure.
+   * @param[in]  pSrc       points to the block of input data.
+   * @param[out] pDst       points to the block of output data.
+   * @param[in]  blockSize  number of samples to process.
+   */
+  void arm_fir_f64(
+  const arm_fir_instance_f64 * S,
+  const float64_t * pSrc,
+        float64_t * pDst,
+        uint32_t blockSize);
+
+  /**
+   * @brief  Initialization function for the floating-point FIR filter.
+   * @param[in,out] S          points to an instance of the floating-point FIR filter structure.
+   * @param[in]     numTaps    Number of filter coefficients in the filter.
+   * @param[in]     pCoeffs    points to the filter coefficients.
+   * @param[in]     pState     points to the state buffer.
+   * @param[in]     blockSize  number of samples that are processed at a time.
+   */
+  void arm_fir_init_f32(
+        arm_fir_instance_f32 * S,
+        uint16_t numTaps,
+  const float32_t * pCoeffs,
+        float32_t * pState,
+        uint32_t blockSize);
+
+  /**
+   * @brief  Initialization function for the floating-point FIR filter.
+   * @param[in,out] S          points to an instance of the floating-point FIR filter structure.
+   * @param[in]     numTaps    Number of filter coefficients in the filter.
+   * @param[in]     pCoeffs    points to the filter coefficients.
+   * @param[in]     pState     points to the state buffer.
+   * @param[in]     blockSize  number of samples that are processed at a time.
+   */
+  void arm_fir_init_f64(
+        arm_fir_instance_f64 * S,
+        uint16_t numTaps,
+  const float64_t * pCoeffs,
+        float64_t * pState,
+        uint32_t blockSize);
+
+  /**
+   * @brief Instance structure for the Q15 Biquad cascade filter.
+   */
+  typedef struct
+  {
+          int8_t numStages;        /**< number of 2nd order stages in the filter.  Overall order is 2*numStages. */
+          q15_t *pState;           /**< Points to the array of state coefficients.  The array is of length 4*numStages. */
+    const q15_t *pCoeffs;          /**< Points to the array of coefficients.  The array is of length 5*numStages. */
+          int8_t postShift;        /**< Additional shift, in bits, applied to each output sample. */
+  } arm_biquad_casd_df1_inst_q15;
+
+  /**
+   * @brief Instance structure for the Q31 Biquad cascade filter.
+   */
+  typedef struct
+  {
+          uint32_t numStages;      /**< number of 2nd order stages in the filter.  Overall order is 2*numStages. */
+          q31_t *pState;           /**< Points to the array of state coefficients.  The array is of length 4*numStages. */
+    const q31_t *pCoeffs;          /**< Points to the array of coefficients.  The array is of length 5*numStages. */
+          uint8_t postShift;       /**< Additional shift, in bits, applied to each output sample. */
+  } arm_biquad_casd_df1_inst_q31;
+
+  /**
+   * @brief Instance structure for the floating-point Biquad cascade filter.
+   */
+  typedef struct
+  {
+          uint32_t numStages;      /**< number of 2nd order stages in the filter.  Overall order is 2*numStages. */
+          float32_t *pState;       /**< Points to the array of state coefficients.  The array is of length 4*numStages. */
+    const float32_t *pCoeffs;      /**< Points to the array of coefficients.  The array is of length 5*numStages. */
+  } arm_biquad_casd_df1_inst_f32;
+
+#if defined(ARM_MATH_MVEF) && !defined(ARM_MATH_AUTOVECTORIZE)
+  /**
+   * @brief Instance structure for the modified Biquad coefs required by vectorized code.
+   */
+  typedef struct
+  {
+      float32_t coeffs[8][4]; /**< Points to the array of modified coefficients.  The array is of length 32. There is one per stage */
+  } arm_biquad_mod_coef_f32;
+#endif 
+
+  /**
+   * @brief Processing function for the Q15 Biquad cascade filter.
+   * @param[in]  S          points to an instance of the Q15 Biquad cascade structure.
+   * @param[in]  pSrc       points to the block of input data.
+   * @param[out] pDst       points to the block of output data.
+   * @param[in]  blockSize  number of samples to process.
+   */
+  void arm_biquad_cascade_df1_q15(
+  const arm_biquad_casd_df1_inst_q15 * S,
+  const q15_t * pSrc,
+        q15_t * pDst,
+        uint32_t blockSize);
+
+  /**
+   * @brief  Initialization function for the Q15 Biquad cascade filter.
+   * @param[in,out] S          points to an instance of the Q15 Biquad cascade structure.
+   * @param[in]     numStages  number of 2nd order stages in the filter.
+   * @param[in]     pCoeffs    points to the filter coefficients.
+   * @param[in]     pState     points to the state buffer.
+   * @param[in]     postShift  Shift to be applied to the output. Varies according to the coefficients format
+   */
+  void arm_biquad_cascade_df1_init_q15(
+        arm_biquad_casd_df1_inst_q15 * S,
+        uint8_t numStages,
+  const q15_t * pCoeffs,
+        q15_t * pState,
+        int8_t postShift);
+
+  /**
+   * @brief Fast but less precise processing function for the Q15 Biquad cascade filter for Cortex-M3 and Cortex-M4.
+   * @param[in]  S          points to an instance of the Q15 Biquad cascade structure.
+   * @param[in]  pSrc       points to the block of input data.
+   * @param[out] pDst       points to the block of output data.
+   * @param[in]  blockSize  number of samples to process.
+   */
+  void arm_biquad_cascade_df1_fast_q15(
+  const arm_biquad_casd_df1_inst_q15 * S,
+  const q15_t * pSrc,
+        q15_t * pDst,
+        uint32_t blockSize);
+
+  /**
+   * @brief Processing function for the Q31 Biquad cascade filter
+   * @param[in]  S          points to an instance of the Q31 Biquad cascade structure.
+   * @param[in]  pSrc       points to the block of input data.
+   * @param[out] pDst       points to the block of output data.
+   * @param[in]  blockSize  number of samples to process.
+   */
+  void arm_biquad_cascade_df1_q31(
+  const arm_biquad_casd_df1_inst_q31 * S,
+  const q31_t * pSrc,
+        q31_t * pDst,
+        uint32_t blockSize);
+
+  /**
+   * @brief Fast but less precise processing function for the Q31 Biquad cascade filter for Cortex-M3 and Cortex-M4.
+   * @param[in]  S          points to an instance of the Q31 Biquad cascade structure.
+   * @param[in]  pSrc       points to the block of input data.
+   * @param[out] pDst       points to the block of output data.
+   * @param[in]  blockSize  number of samples to process.
+   */
+  void arm_biquad_cascade_df1_fast_q31(
+  const arm_biquad_casd_df1_inst_q31 * S,
+  const q31_t * pSrc,
+        q31_t * pDst,
+        uint32_t blockSize);
+
+  /**
+   * @brief  Initialization function for the Q31 Biquad cascade filter.
+   * @param[in,out] S          points to an instance of the Q31 Biquad cascade structure.
+   * @param[in]     numStages  number of 2nd order stages in the filter.
+   * @param[in]     pCoeffs    points to the filter coefficients.
+   * @param[in]     pState     points to the state buffer.
+   * @param[in]     postShift  Shift to be applied to the output. Varies according to the coefficients format
+   */
+  void arm_biquad_cascade_df1_init_q31(
+        arm_biquad_casd_df1_inst_q31 * S,
+        uint8_t numStages,
+  const q31_t * pCoeffs,
+        q31_t * pState,
+        int8_t postShift);
+
+  /**
+   * @brief Processing function for the floating-point Biquad cascade filter.
+   * @param[in]  S          points to an instance of the floating-point Biquad cascade structure.
+   * @param[in]  pSrc       points to the block of input data.
+   * @param[out] pDst       points to the block of output data.
+   * @param[in]  blockSize  number of samples to process.
+   */
+  void arm_biquad_cascade_df1_f32(
+  const arm_biquad_casd_df1_inst_f32 * S,
+  const float32_t * pSrc,
+        float32_t * pDst,
+        uint32_t blockSize);
+
+  /**
+   * @brief  Initialization function for the floating-point Biquad cascade filter.
+   * @param[in,out] S          points to an instance of the floating-point Biquad cascade structure.
+   * @param[in]     numStages  number of 2nd order stages in the filter.
+   * @param[in]     pCoeffs    points to the filter coefficients.
+   * @param[in]     pCoeffsMod points to the modified filter coefficients (only MVE version).
+   * @param[in]     pState     points to the state buffer.
+   */
+#if defined(ARM_MATH_MVEF) && !defined(ARM_MATH_AUTOVECTORIZE)
+  void arm_biquad_cascade_df1_mve_init_f32(
+      arm_biquad_casd_df1_inst_f32 * S,
+      uint8_t numStages,
+      const float32_t * pCoeffs, 
+      arm_biquad_mod_coef_f32 * pCoeffsMod, 
+      float32_t * pState);
+#endif
+  
+  void arm_biquad_cascade_df1_init_f32(
+        arm_biquad_casd_df1_inst_f32 * S,
+        uint8_t numStages,
+  const float32_t * pCoeffs,
+        float32_t * pState);
+
+
+/**
+ * @brief Convolution of floating-point sequences.
+ * @param[in]  pSrcA    points to the first input sequence.
+ * @param[in]  srcALen  length of the first input sequence.
+ * @param[in]  pSrcB    points to the second input sequence.
+ * @param[in]  srcBLen  length of the second input sequence.
+ * @param[out] pDst     points to the location where the output result is written.  Length srcALen+srcBLen-1.
+ */
+  void arm_conv_f32(
+  const float32_t * pSrcA,
+        uint32_t srcALen,
+  const float32_t * pSrcB,
+        uint32_t srcBLen,
+        float32_t * pDst);
+
+
+  /**
+   * @brief Convolution of Q15 sequences.
+   * @param[in]  pSrcA      points to the first input sequence.
+   * @param[in]  srcALen    length of the first input sequence.
+   * @param[in]  pSrcB      points to the second input sequence.
+   * @param[in]  srcBLen    length of the second input sequence.
+   * @param[out] pDst       points to the block of output data  Length srcALen+srcBLen-1.
+   * @param[in]  pScratch1  points to scratch buffer of size max(srcALen, srcBLen) + 2*min(srcALen, srcBLen) - 2.
+   * @param[in]  pScratch2  points to scratch buffer of size min(srcALen, srcBLen).
+   */
+  void arm_conv_opt_q15(
+  const q15_t * pSrcA,
+        uint32_t srcALen,
+  const q15_t * pSrcB,
+        uint32_t srcBLen,
+        q15_t * pDst,
+        q15_t * pScratch1,
+        q15_t * pScratch2);
+
+
+/**
+ * @brief Convolution of Q15 sequences.
+ * @param[in]  pSrcA    points to the first input sequence.
+ * @param[in]  srcALen  length of the first input sequence.
+ * @param[in]  pSrcB    points to the second input sequence.
+ * @param[in]  srcBLen  length of the second input sequence.
+ * @param[out] pDst     points to the location where the output result is written.  Length srcALen+srcBLen-1.
+ */
+  void arm_conv_q15(
+  const q15_t * pSrcA,
+        uint32_t srcALen,
+  const q15_t * pSrcB,
+        uint32_t srcBLen,
+        q15_t * pDst);
+
+
+  /**
+   * @brief Convolution of Q15 sequences (fast version) for Cortex-M3 and Cortex-M4
+   * @param[in]  pSrcA    points to the first input sequence.
+   * @param[in]  srcALen  length of the first input sequence.
+   * @param[in]  pSrcB    points to the second input sequence.
+   * @param[in]  srcBLen  length of the second input sequence.
+   * @param[out] pDst     points to the block of output data  Length srcALen+srcBLen-1.
+   */
+  void arm_conv_fast_q15(
+  const q15_t * pSrcA,
+        uint32_t srcALen,
+  const q15_t * pSrcB,
+        uint32_t srcBLen,
+        q15_t * pDst);
+
+
+  /**
+   * @brief Convolution of Q15 sequences (fast version) for Cortex-M3 and Cortex-M4
+   * @param[in]  pSrcA      points to the first input sequence.
+   * @param[in]  srcALen    length of the first input sequence.
+   * @param[in]  pSrcB      points to the second input sequence.
+   * @param[in]  srcBLen    length of the second input sequence.
+   * @param[out] pDst       points to the block of output data  Length srcALen+srcBLen-1.
+   * @param[in]  pScratch1  points to scratch buffer of size max(srcALen, srcBLen) + 2*min(srcALen, srcBLen) - 2.
+   * @param[in]  pScratch2  points to scratch buffer of size min(srcALen, srcBLen).
+   */
+  void arm_conv_fast_opt_q15(
+  const q15_t * pSrcA,
+        uint32_t srcALen,
+  const q15_t * pSrcB,
+        uint32_t srcBLen,
+        q15_t * pDst,
+        q15_t * pScratch1,
+        q15_t * pScratch2);
+
+
+  /**
+   * @brief Convolution of Q31 sequences.
+   * @param[in]  pSrcA    points to the first input sequence.
+   * @param[in]  srcALen  length of the first input sequence.
+   * @param[in]  pSrcB    points to the second input sequence.
+   * @param[in]  srcBLen  length of the second input sequence.
+   * @param[out] pDst     points to the block of output data  Length srcALen+srcBLen-1.
+   */
+  void arm_conv_q31(
+  const q31_t * pSrcA,
+        uint32_t srcALen,
+  const q31_t * pSrcB,
+        uint32_t srcBLen,
+        q31_t * pDst);
+
+
+  /**
+   * @brief Convolution of Q31 sequences (fast version) for Cortex-M3 and Cortex-M4
+   * @param[in]  pSrcA    points to the first input sequence.
+   * @param[in]  srcALen  length of the first input sequence.
+   * @param[in]  pSrcB    points to the second input sequence.
+   * @param[in]  srcBLen  length of the second input sequence.
+   * @param[out] pDst     points to the block of output data  Length srcALen+srcBLen-1.
+   */
+  void arm_conv_fast_q31(
+  const q31_t * pSrcA,
+        uint32_t srcALen,
+  const q31_t * pSrcB,
+        uint32_t srcBLen,
+        q31_t * pDst);
+
+
+    /**
+   * @brief Convolution of Q7 sequences.
+   * @param[in]  pSrcA      points to the first input sequence.
+   * @param[in]  srcALen    length of the first input sequence.
+   * @param[in]  pSrcB      points to the second input sequence.
+   * @param[in]  srcBLen    length of the second input sequence.
+   * @param[out] pDst       points to the block of output data  Length srcALen+srcBLen-1.
+   * @param[in]  pScratch1  points to scratch buffer(of type q15_t) of size max(srcALen, srcBLen) + 2*min(srcALen, srcBLen) - 2.
+   * @param[in]  pScratch2  points to scratch buffer (of type q15_t) of size min(srcALen, srcBLen).
+   */
+  void arm_conv_opt_q7(
+  const q7_t * pSrcA,
+        uint32_t srcALen,
+  const q7_t * pSrcB,
+        uint32_t srcBLen,
+        q7_t * pDst,
+        q15_t * pScratch1,
+        q15_t * pScratch2);
+
+
+  /**
+   * @brief Convolution of Q7 sequences.
+   * @param[in]  pSrcA    points to the first input sequence.
+   * @param[in]  srcALen  length of the first input sequence.
+   * @param[in]  pSrcB    points to the second input sequence.
+   * @param[in]  srcBLen  length of the second input sequence.
+   * @param[out] pDst     points to the block of output data  Length srcALen+srcBLen-1.
+   */
+  void arm_conv_q7(
+  const q7_t * pSrcA,
+        uint32_t srcALen,
+  const q7_t * pSrcB,
+        uint32_t srcBLen,
+        q7_t * pDst);
+
+
+  /**
+   * @brief Partial convolution of floating-point sequences.
+   * @param[in]  pSrcA       points to the first input sequence.
+   * @param[in]  srcALen     length of the first input sequence.
+   * @param[in]  pSrcB       points to the second input sequence.
+   * @param[in]  srcBLen     length of the second input sequence.
+   * @param[out] pDst        points to the block of output data
+   * @param[in]  firstIndex  is the first output sample to start with.
+   * @param[in]  numPoints   is the number of output points to be computed.
+   * @return  Returns either ARM_MATH_SUCCESS if the function completed correctly or ARM_MATH_ARGUMENT_ERROR if the requested subset is not in the range [0 srcALen+srcBLen-2].
+   */
+  arm_status arm_conv_partial_f32(
+  const float32_t * pSrcA,
+        uint32_t srcALen,
+  const float32_t * pSrcB,
+        uint32_t srcBLen,
+        float32_t * pDst,
+        uint32_t firstIndex,
+        uint32_t numPoints);
+
+
+  /**
+   * @brief Partial convolution of Q15 sequences.
+   * @param[in]  pSrcA       points to the first input sequence.
+   * @param[in]  srcALen     length of the first input sequence.
+   * @param[in]  pSrcB       points to the second input sequence.
+   * @param[in]  srcBLen     length of the second input sequence.
+   * @param[out] pDst        points to the block of output data
+   * @param[in]  firstIndex  is the first output sample to start with.
+   * @param[in]  numPoints   is the number of output points to be computed.
+   * @param[in]  pScratch1   points to scratch buffer of size max(srcALen, srcBLen) + 2*min(srcALen, srcBLen) - 2.
+   * @param[in]  pScratch2   points to scratch buffer of size min(srcALen, srcBLen).
+   * @return  Returns either ARM_MATH_SUCCESS if the function completed correctly or ARM_MATH_ARGUMENT_ERROR if the requested subset is not in the range [0 srcALen+srcBLen-2].
+   */
+  arm_status arm_conv_partial_opt_q15(
+  const q15_t * pSrcA,
+        uint32_t srcALen,
+  const q15_t * pSrcB,
+        uint32_t srcBLen,
+        q15_t * pDst,
+        uint32_t firstIndex,
+        uint32_t numPoints,
+        q15_t * pScratch1,
+        q15_t * pScratch2);
+
+
+  /**
+   * @brief Partial convolution of Q15 sequences.
+   * @param[in]  pSrcA       points to the first input sequence.
+   * @param[in]  srcALen     length of the first input sequence.
+   * @param[in]  pSrcB       points to the second input sequence.
+   * @param[in]  srcBLen     length of the second input sequence.
+   * @param[out] pDst        points to the block of output data
+   * @param[in]  firstIndex  is the first output sample to start with.
+   * @param[in]  numPoints   is the number of output points to be computed.
+   * @return  Returns either ARM_MATH_SUCCESS if the function completed correctly or ARM_MATH_ARGUMENT_ERROR if the requested subset is not in the range [0 srcALen+srcBLen-2].
+   */
+  arm_status arm_conv_partial_q15(
+  const q15_t * pSrcA,
+        uint32_t srcALen,
+  const q15_t * pSrcB,
+        uint32_t srcBLen,
+        q15_t * pDst,
+        uint32_t firstIndex,
+        uint32_t numPoints);
+
+
+  /**
+   * @brief Partial convolution of Q15 sequences (fast version) for Cortex-M3 and Cortex-M4
+   * @param[in]  pSrcA       points to the first input sequence.
+   * @param[in]  srcALen     length of the first input sequence.
+   * @param[in]  pSrcB       points to the second input sequence.
+   * @param[in]  srcBLen     length of the second input sequence.
+   * @param[out] pDst        points to the block of output data
+   * @param[in]  firstIndex  is the first output sample to start with.
+   * @param[in]  numPoints   is the number of output points to be computed.
+   * @return  Returns either ARM_MATH_SUCCESS if the function completed correctly or ARM_MATH_ARGUMENT_ERROR if the requested subset is not in the range [0 srcALen+srcBLen-2].
+   */
+  arm_status arm_conv_partial_fast_q15(
+  const q15_t * pSrcA,
+        uint32_t srcALen,
+  const q15_t * pSrcB,
+        uint32_t srcBLen,
+        q15_t * pDst,
+        uint32_t firstIndex,
+        uint32_t numPoints);
+
+
+  /**
+   * @brief Partial convolution of Q15 sequences (fast version) for Cortex-M3 and Cortex-M4
+   * @param[in]  pSrcA       points to the first input sequence.
+   * @param[in]  srcALen     length of the first input sequence.
+   * @param[in]  pSrcB       points to the second input sequence.
+   * @param[in]  srcBLen     length of the second input sequence.
+   * @param[out] pDst        points to the block of output data
+   * @param[in]  firstIndex  is the first output sample to start with.
+   * @param[in]  numPoints   is the number of output points to be computed.
+   * @param[in]  pScratch1   points to scratch buffer of size max(srcALen, srcBLen) + 2*min(srcALen, srcBLen) - 2.
+   * @param[in]  pScratch2   points to scratch buffer of size min(srcALen, srcBLen).
+   * @return  Returns either ARM_MATH_SUCCESS if the function completed correctly or ARM_MATH_ARGUMENT_ERROR if the requested subset is not in the range [0 srcALen+srcBLen-2].
+   */
+  arm_status arm_conv_partial_fast_opt_q15(
+  const q15_t * pSrcA,
+        uint32_t srcALen,
+  const q15_t * pSrcB,
+        uint32_t srcBLen,
+        q15_t * pDst,
+        uint32_t firstIndex,
+        uint32_t numPoints,
+        q15_t * pScratch1,
+        q15_t * pScratch2);
+
+
+  /**
+   * @brief Partial convolution of Q31 sequences.
+   * @param[in]  pSrcA       points to the first input sequence.
+   * @param[in]  srcALen     length of the first input sequence.
+   * @param[in]  pSrcB       points to the second input sequence.
+   * @param[in]  srcBLen     length of the second input sequence.
+   * @param[out] pDst        points to the block of output data
+   * @param[in]  firstIndex  is the first output sample to start with.
+   * @param[in]  numPoints   is the number of output points to be computed.
+   * @return  Returns either ARM_MATH_SUCCESS if the function completed correctly or ARM_MATH_ARGUMENT_ERROR if the requested subset is not in the range [0 srcALen+srcBLen-2].
+   */
+  arm_status arm_conv_partial_q31(
+  const q31_t * pSrcA,
+        uint32_t srcALen,
+  const q31_t * pSrcB,
+        uint32_t srcBLen,
+        q31_t * pDst,
+        uint32_t firstIndex,
+        uint32_t numPoints);
+
+
+  /**
+   * @brief Partial convolution of Q31 sequences (fast version) for Cortex-M3 and Cortex-M4
+   * @param[in]  pSrcA       points to the first input sequence.
+   * @param[in]  srcALen     length of the first input sequence.
+   * @param[in]  pSrcB       points to the second input sequence.
+   * @param[in]  srcBLen     length of the second input sequence.
+   * @param[out] pDst        points to the block of output data
+   * @param[in]  firstIndex  is the first output sample to start with.
+   * @param[in]  numPoints   is the number of output points to be computed.
+   * @return  Returns either ARM_MATH_SUCCESS if the function completed correctly or ARM_MATH_ARGUMENT_ERROR if the requested subset is not in the range [0 srcALen+srcBLen-2].
+   */
+  arm_status arm_conv_partial_fast_q31(
+  const q31_t * pSrcA,
+        uint32_t srcALen,
+  const q31_t * pSrcB,
+        uint32_t srcBLen,
+        q31_t * pDst,
+        uint32_t firstIndex,
+        uint32_t numPoints);
+
+
+  /**
+   * @brief Partial convolution of Q7 sequences
+   * @param[in]  pSrcA       points to the first input sequence.
+   * @param[in]  srcALen     length of the first input sequence.
+   * @param[in]  pSrcB       points to the second input sequence.
+   * @param[in]  srcBLen     length of the second input sequence.
+   * @param[out] pDst        points to the block of output data
+   * @param[in]  firstIndex  is the first output sample to start with.
+   * @param[in]  numPoints   is the number of output points to be computed.
+   * @param[in]  pScratch1   points to scratch buffer(of type q15_t) of size max(srcALen, srcBLen) + 2*min(srcALen, srcBLen) - 2.
+   * @param[in]  pScratch2   points to scratch buffer (of type q15_t) of size min(srcALen, srcBLen).
+   * @return  Returns either ARM_MATH_SUCCESS if the function completed correctly or ARM_MATH_ARGUMENT_ERROR if the requested subset is not in the range [0 srcALen+srcBLen-2].
+   */
+  arm_status arm_conv_partial_opt_q7(
+  const q7_t * pSrcA,
+        uint32_t srcALen,
+  const q7_t * pSrcB,
+        uint32_t srcBLen,
+        q7_t * pDst,
+        uint32_t firstIndex,
+        uint32_t numPoints,
+        q15_t * pScratch1,
+        q15_t * pScratch2);
+
+
+/**
+   * @brief Partial convolution of Q7 sequences.
+   * @param[in]  pSrcA       points to the first input sequence.
+   * @param[in]  srcALen     length of the first input sequence.
+   * @param[in]  pSrcB       points to the second input sequence.
+   * @param[in]  srcBLen     length of the second input sequence.
+   * @param[out] pDst        points to the block of output data
+   * @param[in]  firstIndex  is the first output sample to start with.
+   * @param[in]  numPoints   is the number of output points to be computed.
+   * @return  Returns either ARM_MATH_SUCCESS if the function completed correctly or ARM_MATH_ARGUMENT_ERROR if the requested subset is not in the range [0 srcALen+srcBLen-2].
+   */
+  arm_status arm_conv_partial_q7(
+  const q7_t * pSrcA,
+        uint32_t srcALen,
+  const q7_t * pSrcB,
+        uint32_t srcBLen,
+        q7_t * pDst,
+        uint32_t firstIndex,
+        uint32_t numPoints);
+
+
+  /**
+   * @brief Instance structure for the Q15 FIR decimator.
+   */
+  typedef struct
+  {
+          uint8_t M;                  /**< decimation factor. */
+          uint16_t numTaps;           /**< number of coefficients in the filter. */
+    const q15_t *pCoeffs;             /**< points to the coefficient array. The array is of length numTaps.*/
+          q15_t *pState;              /**< points to the state variable array. The array is of length numTaps+blockSize-1. */
+  } arm_fir_decimate_instance_q15;
+
+  /**
+   * @brief Instance structure for the Q31 FIR decimator.
+   */
+  typedef struct
+  {
+          uint8_t M;                  /**< decimation factor. */
+          uint16_t numTaps;           /**< number of coefficients in the filter. */
+    const q31_t *pCoeffs;             /**< points to the coefficient array. The array is of length numTaps.*/
+          q31_t *pState;              /**< points to the state variable array. The array is of length numTaps+blockSize-1. */
+  } arm_fir_decimate_instance_q31;
+
+/**
+  @brief Instance structure for floating-point FIR decimator.
+ */
+typedef struct
+  {
+          uint8_t M;                  /**< decimation factor. */
+          uint16_t numTaps;           /**< number of coefficients in the filter. */
+    const float32_t *pCoeffs;         /**< points to the coefficient array. The array is of length numTaps.*/
+          float32_t *pState;          /**< points to the state variable array. The array is of length numTaps+blockSize-1. */
+  } arm_fir_decimate_instance_f32;
+
+
+/**
+  @brief         Processing function for floating-point FIR decimator.
+  @param[in]     S         points to an instance of the floating-point FIR decimator structure
+  @param[in]     pSrc      points to the block of input data
+  @param[out]    pDst      points to the block of output data
+  @param[in]     blockSize number of samples to process
+ */
+void arm_fir_decimate_f32(
+  const arm_fir_decimate_instance_f32 * S,
+  const float32_t * pSrc,
+        float32_t * pDst,
+        uint32_t blockSize);
+
+
+/**
+  @brief         Initialization function for the floating-point FIR decimator.
+  @param[in,out] S          points to an instance of the floating-point FIR decimator structure
+  @param[in]     numTaps    number of coefficients in the filter
+  @param[in]     M          decimation factor
+  @param[in]     pCoeffs    points to the filter coefficients
+  @param[in]     pState     points to the state buffer
+  @param[in]     blockSize  number of input samples to process per call
+  @return        execution status
+                   - \ref ARM_MATH_SUCCESS      : Operation successful
+                   - \ref ARM_MATH_LENGTH_ERROR : <code>blockSize</code> is not a multiple of <code>M</code>
+ */
+arm_status arm_fir_decimate_init_f32(
+        arm_fir_decimate_instance_f32 * S,
+        uint16_t numTaps,
+        uint8_t M,
+  const float32_t * pCoeffs,
+        float32_t * pState,
+        uint32_t blockSize);
+
+
+  /**
+   * @brief Processing function for the Q15 FIR decimator.
+   * @param[in]  S          points to an instance of the Q15 FIR decimator structure.
+   * @param[in]  pSrc       points to the block of input data.
+   * @param[out] pDst       points to the block of output data
+   * @param[in]  blockSize  number of input samples to process per call.
+   */
+  void arm_fir_decimate_q15(
+  const arm_fir_decimate_instance_q15 * S,
+  const q15_t * pSrc,
+        q15_t * pDst,
+        uint32_t blockSize);
+
+
+  /**
+   * @brief Processing function for the Q15 FIR decimator (fast variant) for Cortex-M3 and Cortex-M4.
+   * @param[in]  S          points to an instance of the Q15 FIR decimator structure.
+   * @param[in]  pSrc       points to the block of input data.
+   * @param[out] pDst       points to the block of output data
+   * @param[in]  blockSize  number of input samples to process per call.
+   */
+  void arm_fir_decimate_fast_q15(
+  const arm_fir_decimate_instance_q15 * S,
+  const q15_t * pSrc,
+        q15_t * pDst,
+        uint32_t blockSize);
+
+
+  /**
+   * @brief  Initialization function for the Q15 FIR decimator.
+   * @param[in,out] S          points to an instance of the Q15 FIR decimator structure.
+   * @param[in]     numTaps    number of coefficients in the filter.
+   * @param[in]     M          decimation factor.
+   * @param[in]     pCoeffs    points to the filter coefficients.
+   * @param[in]     pState     points to the state buffer.
+   * @param[in]     blockSize  number of input samples to process per call.
+   * @return    The function returns ARM_MATH_SUCCESS if initialization is successful or ARM_MATH_LENGTH_ERROR if
+   * <code>blockSize</code> is not a multiple of <code>M</code>.
+   */
+  arm_status arm_fir_decimate_init_q15(
+        arm_fir_decimate_instance_q15 * S,
+        uint16_t numTaps,
+        uint8_t M,
+  const q15_t * pCoeffs,
+        q15_t * pState,
+        uint32_t blockSize);
+
+
+  /**
+   * @brief Processing function for the Q31 FIR decimator.
+   * @param[in]  S     points to an instance of the Q31 FIR decimator structure.
+   * @param[in]  pSrc  points to the block of input data.
+   * @param[out] pDst  points to the block of output data
+   * @param[in] blockSize number of input samples to process per call.
+   */
+  void arm_fir_decimate_q31(
+  const arm_fir_decimate_instance_q31 * S,
+  const q31_t * pSrc,
+        q31_t * pDst,
+        uint32_t blockSize);
+
+  /**
+   * @brief Processing function for the Q31 FIR decimator (fast variant) for Cortex-M3 and Cortex-M4.
+   * @param[in]  S          points to an instance of the Q31 FIR decimator structure.
+   * @param[in]  pSrc       points to the block of input data.
+   * @param[out] pDst       points to the block of output data
+   * @param[in]  blockSize  number of input samples to process per call.
+   */
+  void arm_fir_decimate_fast_q31(
+  const arm_fir_decimate_instance_q31 * S,
+  const q31_t * pSrc,
+        q31_t * pDst,
+        uint32_t blockSize);
+
+
+  /**
+   * @brief  Initialization function for the Q31 FIR decimator.
+   * @param[in,out] S          points to an instance of the Q31 FIR decimator structure.
+   * @param[in]     numTaps    number of coefficients in the filter.
+   * @param[in]     M          decimation factor.
+   * @param[in]     pCoeffs    points to the filter coefficients.
+   * @param[in]     pState     points to the state buffer.
+   * @param[in]     blockSize  number of input samples to process per call.
+   * @return    The function returns ARM_MATH_SUCCESS if initialization is successful or ARM_MATH_LENGTH_ERROR if
+   * <code>blockSize</code> is not a multiple of <code>M</code>.
+   */
+  arm_status arm_fir_decimate_init_q31(
+        arm_fir_decimate_instance_q31 * S,
+        uint16_t numTaps,
+        uint8_t M,
+  const q31_t * pCoeffs,
+        q31_t * pState,
+        uint32_t blockSize);
+
+
+  /**
+   * @brief Instance structure for the Q15 FIR interpolator.
+   */
+  typedef struct
+  {
+        uint8_t L;                      /**< upsample factor. */
+        uint16_t phaseLength;           /**< length of each polyphase filter component. */
+  const q15_t *pCoeffs;                 /**< points to the coefficient array. The array is of length L*phaseLength. */
+        q15_t *pState;                  /**< points to the state variable array. The array is of length blockSize+phaseLength-1. */
+  } arm_fir_interpolate_instance_q15;
+
+  /**
+   * @brief Instance structure for the Q31 FIR interpolator.
+   */
+  typedef struct
+  {
+        uint8_t L;                      /**< upsample factor. */
+        uint16_t phaseLength;           /**< length of each polyphase filter component. */
+  const q31_t *pCoeffs;                 /**< points to the coefficient array. The array is of length L*phaseLength. */
+        q31_t *pState;                  /**< points to the state variable array. The array is of length blockSize+phaseLength-1. */
+  } arm_fir_interpolate_instance_q31;
+
+  /**
+   * @brief Instance structure for the floating-point FIR interpolator.
+   */
+  typedef struct
+  {
+        uint8_t L;                     /**< upsample factor. */
+        uint16_t phaseLength;          /**< length of each polyphase filter component. */
+  const float32_t *pCoeffs;            /**< points to the coefficient array. The array is of length L*phaseLength. */
+        float32_t *pState;             /**< points to the state variable array. The array is of length phaseLength+numTaps-1. */
+  } arm_fir_interpolate_instance_f32;
+
+
+  /**
+   * @brief Processing function for the Q15 FIR interpolator.
+   * @param[in]  S          points to an instance of the Q15 FIR interpolator structure.
+   * @param[in]  pSrc       points to the block of input data.
+   * @param[out] pDst       points to the block of output data.
+   * @param[in]  blockSize  number of input samples to process per call.
+   */
+  void arm_fir_interpolate_q15(
+  const arm_fir_interpolate_instance_q15 * S,
+  const q15_t * pSrc,
+        q15_t * pDst,
+        uint32_t blockSize);
+
+
+  /**
+   * @brief  Initialization function for the Q15 FIR interpolator.
+   * @param[in,out] S          points to an instance of the Q15 FIR interpolator structure.
+   * @param[in]     L          upsample factor.
+   * @param[in]     numTaps    number of filter coefficients in the filter.
+   * @param[in]     pCoeffs    points to the filter coefficient buffer.
+   * @param[in]     pState     points to the state buffer.
+   * @param[in]     blockSize  number of input samples to process per call.
+   * @return        The function returns ARM_MATH_SUCCESS if initialization is successful or ARM_MATH_LENGTH_ERROR if
+   * the filter length <code>numTaps</code> is not a multiple of the interpolation factor <code>L</code>.
+   */
+  arm_status arm_fir_interpolate_init_q15(
+        arm_fir_interpolate_instance_q15 * S,
+        uint8_t L,
+        uint16_t numTaps,
+  const q15_t * pCoeffs,
+        q15_t * pState,
+        uint32_t blockSize);
+
+
+  /**
+   * @brief Processing function for the Q31 FIR interpolator.
+   * @param[in]  S          points to an instance of the Q15 FIR interpolator structure.
+   * @param[in]  pSrc       points to the block of input data.
+   * @param[out] pDst       points to the block of output data.
+   * @param[in]  blockSize  number of input samples to process per call.
+   */
+  void arm_fir_interpolate_q31(
+  const arm_fir_interpolate_instance_q31 * S,
+  const q31_t * pSrc,
+        q31_t * pDst,
+        uint32_t blockSize);
+
+
+  /**
+   * @brief  Initialization function for the Q31 FIR interpolator.
+   * @param[in,out] S          points to an instance of the Q31 FIR interpolator structure.
+   * @param[in]     L          upsample factor.
+   * @param[in]     numTaps    number of filter coefficients in the filter.
+   * @param[in]     pCoeffs    points to the filter coefficient buffer.
+   * @param[in]     pState     points to the state buffer.
+   * @param[in]     blockSize  number of input samples to process per call.
+   * @return        The function returns ARM_MATH_SUCCESS if initialization is successful or ARM_MATH_LENGTH_ERROR if
+   * the filter length <code>numTaps</code> is not a multiple of the interpolation factor <code>L</code>.
+   */
+  arm_status arm_fir_interpolate_init_q31(
+        arm_fir_interpolate_instance_q31 * S,
+        uint8_t L,
+        uint16_t numTaps,
+  const q31_t * pCoeffs,
+        q31_t * pState,
+        uint32_t blockSize);
+
+
+  /**
+   * @brief Processing function for the floating-point FIR interpolator.
+   * @param[in]  S          points to an instance of the floating-point FIR interpolator structure.
+   * @param[in]  pSrc       points to the block of input data.
+   * @param[out] pDst       points to the block of output data.
+   * @param[in]  blockSize  number of input samples to process per call.
+   */
+  void arm_fir_interpolate_f32(
+  const arm_fir_interpolate_instance_f32 * S,
+  const float32_t * pSrc,
+        float32_t * pDst,
+        uint32_t blockSize);
+
+
+  /**
+   * @brief  Initialization function for the floating-point FIR interpolator.
+   * @param[in,out] S          points to an instance of the floating-point FIR interpolator structure.
+   * @param[in]     L          upsample factor.
+   * @param[in]     numTaps    number of filter coefficients in the filter.
+   * @param[in]     pCoeffs    points to the filter coefficient buffer.
+   * @param[in]     pState     points to the state buffer.
+   * @param[in]     blockSize  number of input samples to process per call.
+   * @return        The function returns ARM_MATH_SUCCESS if initialization is successful or ARM_MATH_LENGTH_ERROR if
+   * the filter length <code>numTaps</code> is not a multiple of the interpolation factor <code>L</code>.
+   */
+  arm_status arm_fir_interpolate_init_f32(
+        arm_fir_interpolate_instance_f32 * S,
+        uint8_t L,
+        uint16_t numTaps,
+  const float32_t * pCoeffs,
+        float32_t * pState,
+        uint32_t blockSize);
+
+
+  /**
+   * @brief Instance structure for the high precision Q31 Biquad cascade filter.
+   */
+  typedef struct
+  {
+          uint8_t numStages;       /**< number of 2nd order stages in the filter.  Overall order is 2*numStages. */
+          q63_t *pState;           /**< points to the array of state coefficients.  The array is of length 4*numStages. */
+    const q31_t *pCoeffs;          /**< points to the array of coefficients.  The array is of length 5*numStages. */
+          uint8_t postShift;       /**< additional shift, in bits, applied to each output sample. */
+  } arm_biquad_cas_df1_32x64_ins_q31;
+
+
+  /**
+   * @param[in]  S          points to an instance of the high precision Q31 Biquad cascade filter structure.
+   * @param[in]  pSrc       points to the block of input data.
+   * @param[out] pDst       points to the block of output data
+   * @param[in]  blockSize  number of samples to process.
+   */
+  void arm_biquad_cas_df1_32x64_q31(
+  const arm_biquad_cas_df1_32x64_ins_q31 * S,
+  const q31_t * pSrc,
+        q31_t * pDst,
+        uint32_t blockSize);
+
+
+  /**
+   * @param[in,out] S          points to an instance of the high precision Q31 Biquad cascade filter structure.
+   * @param[in]     numStages  number of 2nd order stages in the filter.
+   * @param[in]     pCoeffs    points to the filter coefficients.
+   * @param[in]     pState     points to the state buffer.
+   * @param[in]     postShift  shift to be applied to the output. Varies according to the coefficients format
+   */
+  void arm_biquad_cas_df1_32x64_init_q31(
+        arm_biquad_cas_df1_32x64_ins_q31 * S,
+        uint8_t numStages,
+  const q31_t * pCoeffs,
+        q63_t * pState,
+        uint8_t postShift);
+
+
+  /**
+   * @brief Instance structure for the floating-point transposed direct form II Biquad cascade filter.
+   */
+  typedef struct
+  {
+          uint8_t numStages;         /**< number of 2nd order stages in the filter.  Overall order is 2*numStages. */
+          float32_t *pState;         /**< points to the array of state coefficients.  The array is of length 2*numStages. */
+    const float32_t *pCoeffs;        /**< points to the array of coefficients.  The array is of length 5*numStages. */
+  } arm_biquad_cascade_df2T_instance_f32;
+
+  /**
+   * @brief Instance structure for the floating-point transposed direct form II Biquad cascade filter.
+   */
+  typedef struct
+  {
+          uint8_t numStages;         /**< number of 2nd order stages in the filter.  Overall order is 2*numStages. */
+          float32_t *pState;         /**< points to the array of state coefficients.  The array is of length 4*numStages. */
+    const float32_t *pCoeffs;        /**< points to the array of coefficients.  The array is of length 5*numStages. */
+  } arm_biquad_cascade_stereo_df2T_instance_f32;
+
+  /**
+   * @brief Instance structure for the floating-point transposed direct form II Biquad cascade filter.
+   */
+  typedef struct
+  {
+          uint8_t numStages;         /**< number of 2nd order stages in the filter.  Overall order is 2*numStages. */
+          float64_t *pState;         /**< points to the array of state coefficients.  The array is of length 2*numStages. */
+    const float64_t *pCoeffs;        /**< points to the array of coefficients.  The array is of length 5*numStages. */
+  } arm_biquad_cascade_df2T_instance_f64;
+
+
+  /**
+   * @brief Processing function for the floating-point transposed direct form II Biquad cascade filter.
+   * @param[in]  S          points to an instance of the filter data structure.
+   * @param[in]  pSrc       points to the block of input data.
+   * @param[out] pDst       points to the block of output data
+   * @param[in]  blockSize  number of samples to process.
+   */
+  void arm_biquad_cascade_df2T_f32(
+  const arm_biquad_cascade_df2T_instance_f32 * S,
+  const float32_t * pSrc,
+        float32_t * pDst,
+        uint32_t blockSize);
+
+
+  /**
+   * @brief Processing function for the floating-point transposed direct form II Biquad cascade filter. 2 channels
+   * @param[in]  S          points to an instance of the filter data structure.
+   * @param[in]  pSrc       points to the block of input data.
+   * @param[out] pDst       points to the block of output data
+   * @param[in]  blockSize  number of samples to process.
+   */
+  void arm_biquad_cascade_stereo_df2T_f32(
+  const arm_biquad_cascade_stereo_df2T_instance_f32 * S,
+  const float32_t * pSrc,
+        float32_t * pDst,
+        uint32_t blockSize);
+
+
+  /**
+   * @brief Processing function for the floating-point transposed direct form II Biquad cascade filter.
+   * @param[in]  S          points to an instance of the filter data structure.
+   * @param[in]  pSrc       points to the block of input data.
+   * @param[out] pDst       points to the block of output data
+   * @param[in]  blockSize  number of samples to process.
+   */
+  void arm_biquad_cascade_df2T_f64(
+  const arm_biquad_cascade_df2T_instance_f64 * S,
+  const float64_t * pSrc,
+        float64_t * pDst,
+        uint32_t blockSize);
+
+
+#if defined(ARM_MATH_NEON) 
+/**
+  @brief         Compute new coefficient arrays for use in vectorized filter (Neon only).
+  @param[in]     numStages         number of 2nd order stages in the filter.
+  @param[in]     pCoeffs           points to the original filter coefficients.
+  @param[in]     pComputedCoeffs   points to the new computed coefficients for the vectorized version.
+  @return        none
+*/
+void arm_biquad_cascade_df2T_compute_coefs_f32(
+  uint8_t numStages,
+  const float32_t * pCoeffs,
+  float32_t * pComputedCoeffs);
+#endif
+  /**
+   * @brief  Initialization function for the floating-point transposed direct form II Biquad cascade filter.
+   * @param[in,out] S          points to an instance of the filter data structure.
+   * @param[in]     numStages  number of 2nd order stages in the filter.
+   * @param[in]     pCoeffs    points to the filter coefficients.
+   * @param[in]     pState     points to the state buffer.
+   */
+  void arm_biquad_cascade_df2T_init_f32(
+        arm_biquad_cascade_df2T_instance_f32 * S,
+        uint8_t numStages,
+  const float32_t * pCoeffs,
+        float32_t * pState);
+
+
+  /**
+   * @brief  Initialization function for the floating-point transposed direct form II Biquad cascade filter.
+   * @param[in,out] S          points to an instance of the filter data structure.
+   * @param[in]     numStages  number of 2nd order stages in the filter.
+   * @param[in]     pCoeffs    points to the filter coefficients.
+   * @param[in]     pState     points to the state buffer.
+   */
+  void arm_biquad_cascade_stereo_df2T_init_f32(
+        arm_biquad_cascade_stereo_df2T_instance_f32 * S,
+        uint8_t numStages,
+  const float32_t * pCoeffs,
+        float32_t * pState);
+
+
+  /**
+   * @brief  Initialization function for the floating-point transposed direct form II Biquad cascade filter.
+   * @param[in,out] S          points to an instance of the filter data structure.
+   * @param[in]     numStages  number of 2nd order stages in the filter.
+   * @param[in]     pCoeffs    points to the filter coefficients.
+   * @param[in]     pState     points to the state buffer.
+   */
+  void arm_biquad_cascade_df2T_init_f64(
+        arm_biquad_cascade_df2T_instance_f64 * S,
+        uint8_t numStages,
+        const float64_t * pCoeffs,
+        float64_t * pState);
+
+
+  /**
+   * @brief Instance structure for the Q15 FIR lattice filter.
+   */
+  typedef struct
+  {
+          uint16_t numStages;                  /**< number of filter stages. */
+          q15_t *pState;                       /**< points to the state variable array. The array is of length numStages. */
+    const q15_t *pCoeffs;                      /**< points to the coefficient array. The array is of length numStages. */
+  } arm_fir_lattice_instance_q15;
+
+  /**
+   * @brief Instance structure for the Q31 FIR lattice filter.
+   */
+  typedef struct
+  {
+          uint16_t numStages;                  /**< number of filter stages. */
+          q31_t *pState;                       /**< points to the state variable array. The array is of length numStages. */
+    const q31_t *pCoeffs;                      /**< points to the coefficient array. The array is of length numStages. */
+  } arm_fir_lattice_instance_q31;
+
+  /**
+   * @brief Instance structure for the floating-point FIR lattice filter.
+   */
+  typedef struct
+  {
+          uint16_t numStages;                  /**< number of filter stages. */
+          float32_t *pState;                   /**< points to the state variable array. The array is of length numStages. */
+    const float32_t *pCoeffs;                  /**< points to the coefficient array. The array is of length numStages. */
+  } arm_fir_lattice_instance_f32;
+
+
+  /**
+   * @brief Initialization function for the Q15 FIR lattice filter.
+   * @param[in] S          points to an instance of the Q15 FIR lattice structure.
+   * @param[in] numStages  number of filter stages.
+   * @param[in] pCoeffs    points to the coefficient buffer.  The array is of length numStages.
+   * @param[in] pState     points to the state buffer.  The array is of length numStages.
+   */
+  void arm_fir_lattice_init_q15(
+        arm_fir_lattice_instance_q15 * S,
+        uint16_t numStages,
+  const q15_t * pCoeffs,
+        q15_t * pState);
+
+
+  /**
+   * @brief Processing function for the Q15 FIR lattice filter.
+   * @param[in]  S          points to an instance of the Q15 FIR lattice structure.
+   * @param[in]  pSrc       points to the block of input data.
+   * @param[out] pDst       points to the block of output data.
+   * @param[in]  blockSize  number of samples to process.
+   */
+  void arm_fir_lattice_q15(
+  const arm_fir_lattice_instance_q15 * S,
+  const q15_t * pSrc,
+        q15_t * pDst,
+        uint32_t blockSize);
+
+
+  /**
+   * @brief Initialization function for the Q31 FIR lattice filter.
+   * @param[in] S          points to an instance of the Q31 FIR lattice structure.
+   * @param[in] numStages  number of filter stages.
+   * @param[in] pCoeffs    points to the coefficient buffer.  The array is of length numStages.
+   * @param[in] pState     points to the state buffer.   The array is of length numStages.
+   */
+  void arm_fir_lattice_init_q31(
+        arm_fir_lattice_instance_q31 * S,
+        uint16_t numStages,
+  const q31_t * pCoeffs,
+        q31_t * pState);
+
+
+  /**
+   * @brief Processing function for the Q31 FIR lattice filter.
+   * @param[in]  S          points to an instance of the Q31 FIR lattice structure.
+   * @param[in]  pSrc       points to the block of input data.
+   * @param[out] pDst       points to the block of output data
+   * @param[in]  blockSize  number of samples to process.
+   */
+  void arm_fir_lattice_q31(
+  const arm_fir_lattice_instance_q31 * S,
+  const q31_t * pSrc,
+        q31_t * pDst,
+        uint32_t blockSize);
+
+
+/**
+ * @brief Initialization function for the floating-point FIR lattice filter.
+ * @param[in] S          points to an instance of the floating-point FIR lattice structure.
+ * @param[in] numStages  number of filter stages.
+ * @param[in] pCoeffs    points to the coefficient buffer.  The array is of length numStages.
+ * @param[in] pState     points to the state buffer.  The array is of length numStages.
+ */
+  void arm_fir_lattice_init_f32(
+        arm_fir_lattice_instance_f32 * S,
+        uint16_t numStages,
+  const float32_t * pCoeffs,
+        float32_t * pState);
+
+
+  /**
+   * @brief Processing function for the floating-point FIR lattice filter.
+   * @param[in]  S          points to an instance of the floating-point FIR lattice structure.
+   * @param[in]  pSrc       points to the block of input data.
+   * @param[out] pDst       points to the block of output data
+   * @param[in]  blockSize  number of samples to process.
+   */
+  void arm_fir_lattice_f32(
+  const arm_fir_lattice_instance_f32 * S,
+  const float32_t * pSrc,
+        float32_t * pDst,
+        uint32_t blockSize);
+
+
+  /**
+   * @brief Instance structure for the Q15 IIR lattice filter.
+   */
+  typedef struct
+  {
+          uint16_t numStages;                  /**< number of stages in the filter. */
+          q15_t *pState;                       /**< points to the state variable array. The array is of length numStages+blockSize. */
+          q15_t *pkCoeffs;                     /**< points to the reflection coefficient array. The array is of length numStages. */
+          q15_t *pvCoeffs;                     /**< points to the ladder coefficient array. The array is of length numStages+1. */
+  } arm_iir_lattice_instance_q15;
+
+  /**
+   * @brief Instance structure for the Q31 IIR lattice filter.
+   */
+  typedef struct
+  {
+          uint16_t numStages;                  /**< number of stages in the filter. */
+          q31_t *pState;                       /**< points to the state variable array. The array is of length numStages+blockSize. */
+          q31_t *pkCoeffs;                     /**< points to the reflection coefficient array. The array is of length numStages. */
+          q31_t *pvCoeffs;                     /**< points to the ladder coefficient array. The array is of length numStages+1. */
+  } arm_iir_lattice_instance_q31;
+
+  /**
+   * @brief Instance structure for the floating-point IIR lattice filter.
+   */
+  typedef struct
+  {
+          uint16_t numStages;                  /**< number of stages in the filter. */
+          float32_t *pState;                   /**< points to the state variable array. The array is of length numStages+blockSize. */
+          float32_t *pkCoeffs;                 /**< points to the reflection coefficient array. The array is of length numStages. */
+          float32_t *pvCoeffs;                 /**< points to the ladder coefficient array. The array is of length numStages+1. */
+  } arm_iir_lattice_instance_f32;
+
+
+  /**
+   * @brief Processing function for the floating-point IIR lattice filter.
+   * @param[in]  S          points to an instance of the floating-point IIR lattice structure.
+   * @param[in]  pSrc       points to the block of input data.
+   * @param[out] pDst       points to the block of output data.
+   * @param[in]  blockSize  number of samples to process.
+   */
+  void arm_iir_lattice_f32(
+  const arm_iir_lattice_instance_f32 * S,
+  const float32_t * pSrc,
+        float32_t * pDst,
+        uint32_t blockSize);
+
+
+  /**
+   * @brief Initialization function for the floating-point IIR lattice filter.
+   * @param[in] S          points to an instance of the floating-point IIR lattice structure.
+   * @param[in] numStages  number of stages in the filter.
+   * @param[in] pkCoeffs   points to the reflection coefficient buffer.  The array is of length numStages.
+   * @param[in] pvCoeffs   points to the ladder coefficient buffer.  The array is of length numStages+1.
+   * @param[in] pState     points to the state buffer.  The array is of length numStages+blockSize-1.
+   * @param[in] blockSize  number of samples to process.
+   */
+  void arm_iir_lattice_init_f32(
+        arm_iir_lattice_instance_f32 * S,
+        uint16_t numStages,
+        float32_t * pkCoeffs,
+        float32_t * pvCoeffs,
+        float32_t * pState,
+        uint32_t blockSize);
+
+
+  /**
+   * @brief Processing function for the Q31 IIR lattice filter.
+   * @param[in]  S          points to an instance of the Q31 IIR lattice structure.
+   * @param[in]  pSrc       points to the block of input data.
+   * @param[out] pDst       points to the block of output data.
+   * @param[in]  blockSize  number of samples to process.
+   */
+  void arm_iir_lattice_q31(
+  const arm_iir_lattice_instance_q31 * S,
+  const q31_t * pSrc,
+        q31_t * pDst,
+        uint32_t blockSize);
+
+
+  /**
+   * @brief Initialization function for the Q31 IIR lattice filter.
+   * @param[in] S          points to an instance of the Q31 IIR lattice structure.
+   * @param[in] numStages  number of stages in the filter.
+   * @param[in] pkCoeffs   points to the reflection coefficient buffer.  The array is of length numStages.
+   * @param[in] pvCoeffs   points to the ladder coefficient buffer.  The array is of length numStages+1.
+   * @param[in] pState     points to the state buffer.  The array is of length numStages+blockSize.
+   * @param[in] blockSize  number of samples to process.
+   */
+  void arm_iir_lattice_init_q31(
+        arm_iir_lattice_instance_q31 * S,
+        uint16_t numStages,
+        q31_t * pkCoeffs,
+        q31_t * pvCoeffs,
+        q31_t * pState,
+        uint32_t blockSize);
+
+
+  /**
+   * @brief Processing function for the Q15 IIR lattice filter.
+   * @param[in]  S          points to an instance of the Q15 IIR lattice structure.
+   * @param[in]  pSrc       points to the block of input data.
+   * @param[out] pDst       points to the block of output data.
+   * @param[in]  blockSize  number of samples to process.
+   */
+  void arm_iir_lattice_q15(
+  const arm_iir_lattice_instance_q15 * S,
+  const q15_t * pSrc,
+        q15_t * pDst,
+        uint32_t blockSize);
+
+
+/**
+ * @brief Initialization function for the Q15 IIR lattice filter.
+ * @param[in] S          points to an instance of the fixed-point Q15 IIR lattice structure.
+ * @param[in] numStages  number of stages in the filter.
+ * @param[in] pkCoeffs   points to reflection coefficient buffer.  The array is of length numStages.
+ * @param[in] pvCoeffs   points to ladder coefficient buffer.  The array is of length numStages+1.
+ * @param[in] pState     points to state buffer.  The array is of length numStages+blockSize.
+ * @param[in] blockSize  number of samples to process per call.
+ */
+  void arm_iir_lattice_init_q15(
+        arm_iir_lattice_instance_q15 * S,
+        uint16_t numStages,
+        q15_t * pkCoeffs,
+        q15_t * pvCoeffs,
+        q15_t * pState,
+        uint32_t blockSize);
+
+
+  /**
+   * @brief Instance structure for the floating-point LMS filter.
+   */
+  typedef struct
+  {
+          uint16_t numTaps;    /**< number of coefficients in the filter. */
+          float32_t *pState;   /**< points to the state variable array. The array is of length numTaps+blockSize-1. */
+          float32_t *pCoeffs;  /**< points to the coefficient array. The array is of length numTaps. */
+          float32_t mu;        /**< step size that controls filter coefficient updates. */
+  } arm_lms_instance_f32;
+
+
+  /**
+   * @brief Processing function for floating-point LMS filter.
+   * @param[in]  S          points to an instance of the floating-point LMS filter structure.
+   * @param[in]  pSrc       points to the block of input data.
+   * @param[in]  pRef       points to the block of reference data.
+   * @param[out] pOut       points to the block of output data.
+   * @param[out] pErr       points to the block of error data.
+   * @param[in]  blockSize  number of samples to process.
+   */
+  void arm_lms_f32(
+  const arm_lms_instance_f32 * S,
+  const float32_t * pSrc,
+        float32_t * pRef,
+        float32_t * pOut,
+        float32_t * pErr,
+        uint32_t blockSize);
+
+
+  /**
+   * @brief Initialization function for floating-point LMS filter.
+   * @param[in] S          points to an instance of the floating-point LMS filter structure.
+   * @param[in] numTaps    number of filter coefficients.
+   * @param[in] pCoeffs    points to the coefficient buffer.
+   * @param[in] pState     points to state buffer.
+   * @param[in] mu         step size that controls filter coefficient updates.
+   * @param[in] blockSize  number of samples to process.
+   */
+  void arm_lms_init_f32(
+        arm_lms_instance_f32 * S,
+        uint16_t numTaps,
+        float32_t * pCoeffs,
+        float32_t * pState,
+        float32_t mu,
+        uint32_t blockSize);
+
+
+  /**
+   * @brief Instance structure for the Q15 LMS filter.
+   */
+  typedef struct
+  {
+          uint16_t numTaps;    /**< number of coefficients in the filter. */
+          q15_t *pState;       /**< points to the state variable array. The array is of length numTaps+blockSize-1. */
+          q15_t *pCoeffs;      /**< points to the coefficient array. The array is of length numTaps. */
+          q15_t mu;            /**< step size that controls filter coefficient updates. */
+          uint32_t postShift;  /**< bit shift applied to coefficients. */
+  } arm_lms_instance_q15;
+
+
+  /**
+   * @brief Initialization function for the Q15 LMS filter.
+   * @param[in] S          points to an instance of the Q15 LMS filter structure.
+   * @param[in] numTaps    number of filter coefficients.
+   * @param[in] pCoeffs    points to the coefficient buffer.
+   * @param[in] pState     points to the state buffer.
+   * @param[in] mu         step size that controls filter coefficient updates.
+   * @param[in] blockSize  number of samples to process.
+   * @param[in] postShift  bit shift applied to coefficients.
+   */
+  void arm_lms_init_q15(
+        arm_lms_instance_q15 * S,
+        uint16_t numTaps,
+        q15_t * pCoeffs,
+        q15_t * pState,
+        q15_t mu,
+        uint32_t blockSize,
+        uint32_t postShift);
+
+
+  /**
+   * @brief Processing function for Q15 LMS filter.
+   * @param[in]  S          points to an instance of the Q15 LMS filter structure.
+   * @param[in]  pSrc       points to the block of input data.
+   * @param[in]  pRef       points to the block of reference data.
+   * @param[out] pOut       points to the block of output data.
+   * @param[out] pErr       points to the block of error data.
+   * @param[in]  blockSize  number of samples to process.
+   */
+  void arm_lms_q15(
+  const arm_lms_instance_q15 * S,
+  const q15_t * pSrc,
+        q15_t * pRef,
+        q15_t * pOut,
+        q15_t * pErr,
+        uint32_t blockSize);
+
+
+  /**
+   * @brief Instance structure for the Q31 LMS filter.
+   */
+  typedef struct
+  {
+          uint16_t numTaps;    /**< number of coefficients in the filter. */
+          q31_t *pState;       /**< points to the state variable array. The array is of length numTaps+blockSize-1. */
+          q31_t *pCoeffs;      /**< points to the coefficient array. The array is of length numTaps. */
+          q31_t mu;            /**< step size that controls filter coefficient updates. */
+          uint32_t postShift;  /**< bit shift applied to coefficients. */
+  } arm_lms_instance_q31;
+
+
+  /**
+   * @brief Processing function for Q31 LMS filter.
+   * @param[in]  S          points to an instance of the Q15 LMS filter structure.
+   * @param[in]  pSrc       points to the block of input data.
+   * @param[in]  pRef       points to the block of reference data.
+   * @param[out] pOut       points to the block of output data.
+   * @param[out] pErr       points to the block of error data.
+   * @param[in]  blockSize  number of samples to process.
+   */
+  void arm_lms_q31(
+  const arm_lms_instance_q31 * S,
+  const q31_t * pSrc,
+        q31_t * pRef,
+        q31_t * pOut,
+        q31_t * pErr,
+        uint32_t blockSize);
+
+
+  /**
+   * @brief Initialization function for Q31 LMS filter.
+   * @param[in] S          points to an instance of the Q31 LMS filter structure.
+   * @param[in] numTaps    number of filter coefficients.
+   * @param[in] pCoeffs    points to coefficient buffer.
+   * @param[in] pState     points to state buffer.
+   * @param[in] mu         step size that controls filter coefficient updates.
+   * @param[in] blockSize  number of samples to process.
+   * @param[in] postShift  bit shift applied to coefficients.
+   */
+  void arm_lms_init_q31(
+        arm_lms_instance_q31 * S,
+        uint16_t numTaps,
+        q31_t * pCoeffs,
+        q31_t * pState,
+        q31_t mu,
+        uint32_t blockSize,
+        uint32_t postShift);
+
+
+  /**
+   * @brief Instance structure for the floating-point normalized LMS filter.
+   */
+  typedef struct
+  {
+          uint16_t numTaps;     /**< number of coefficients in the filter. */
+          float32_t *pState;    /**< points to the state variable array. The array is of length numTaps+blockSize-1. */
+          float32_t *pCoeffs;   /**< points to the coefficient array. The array is of length numTaps. */
+          float32_t mu;         /**< step size that control filter coefficient updates. */
+          float32_t energy;     /**< saves previous frame energy. */
+          float32_t x0;         /**< saves previous input sample. */
+  } arm_lms_norm_instance_f32;
+
+
+  /**
+   * @brief Processing function for floating-point normalized LMS filter.
+   * @param[in]  S          points to an instance of the floating-point normalized LMS filter structure.
+   * @param[in]  pSrc       points to the block of input data.
+   * @param[in]  pRef       points to the block of reference data.
+   * @param[out] pOut       points to the block of output data.
+   * @param[out] pErr       points to the block of error data.
+   * @param[in]  blockSize  number of samples to process.
+   */
+  void arm_lms_norm_f32(
+        arm_lms_norm_instance_f32 * S,
+  const float32_t * pSrc,
+        float32_t * pRef,
+        float32_t * pOut,
+        float32_t * pErr,
+        uint32_t blockSize);
+
+
+  /**
+   * @brief Initialization function for floating-point normalized LMS filter.
+   * @param[in] S          points to an instance of the floating-point LMS filter structure.
+   * @param[in] numTaps    number of filter coefficients.
+   * @param[in] pCoeffs    points to coefficient buffer.
+   * @param[in] pState     points to state buffer.
+   * @param[in] mu         step size that controls filter coefficient updates.
+   * @param[in] blockSize  number of samples to process.
+   */
+  void arm_lms_norm_init_f32(
+        arm_lms_norm_instance_f32 * S,
+        uint16_t numTaps,
+        float32_t * pCoeffs,
+        float32_t * pState,
+        float32_t mu,
+        uint32_t blockSize);
+
+
+  /**
+   * @brief Instance structure for the Q31 normalized LMS filter.
+   */
+  typedef struct
+  {
+          uint16_t numTaps;     /**< number of coefficients in the filter. */
+          q31_t *pState;        /**< points to the state variable array. The array is of length numTaps+blockSize-1. */
+          q31_t *pCoeffs;       /**< points to the coefficient array. The array is of length numTaps. */
+          q31_t mu;             /**< step size that controls filter coefficient updates. */
+          uint8_t postShift;    /**< bit shift applied to coefficients. */
+    const q31_t *recipTable;    /**< points to the reciprocal initial value table. */
+          q31_t energy;         /**< saves previous frame energy. */
+          q31_t x0;             /**< saves previous input sample. */
+  } arm_lms_norm_instance_q31;
+
+
+  /**
+   * @brief Processing function for Q31 normalized LMS filter.
+   * @param[in]  S          points to an instance of the Q31 normalized LMS filter structure.
+   * @param[in]  pSrc       points to the block of input data.
+   * @param[in]  pRef       points to the block of reference data.
+   * @param[out] pOut       points to the block of output data.
+   * @param[out] pErr       points to the block of error data.
+   * @param[in]  blockSize  number of samples to process.
+   */
+  void arm_lms_norm_q31(
+        arm_lms_norm_instance_q31 * S,
+  const q31_t * pSrc,
+        q31_t * pRef,
+        q31_t * pOut,
+        q31_t * pErr,
+        uint32_t blockSize);
+
+
+  /**
+   * @brief Initialization function for Q31 normalized LMS filter.
+   * @param[in] S          points to an instance of the Q31 normalized LMS filter structure.
+   * @param[in] numTaps    number of filter coefficients.
+   * @param[in] pCoeffs    points to coefficient buffer.
+   * @param[in] pState     points to state buffer.
+   * @param[in] mu         step size that controls filter coefficient updates.
+   * @param[in] blockSize  number of samples to process.
+   * @param[in] postShift  bit shift applied to coefficients.
+   */
+  void arm_lms_norm_init_q31(
+        arm_lms_norm_instance_q31 * S,
+        uint16_t numTaps,
+        q31_t * pCoeffs,
+        q31_t * pState,
+        q31_t mu,
+        uint32_t blockSize,
+        uint8_t postShift);
+
+
+  /**
+   * @brief Instance structure for the Q15 normalized LMS filter.
+   */
+  typedef struct
+  {
+          uint16_t numTaps;     /**< Number of coefficients in the filter. */
+          q15_t *pState;        /**< points to the state variable array. The array is of length numTaps+blockSize-1. */
+          q15_t *pCoeffs;       /**< points to the coefficient array. The array is of length numTaps. */
+          q15_t mu;             /**< step size that controls filter coefficient updates. */
+          uint8_t postShift;    /**< bit shift applied to coefficients. */
+    const q15_t *recipTable;    /**< Points to the reciprocal initial value table. */
+          q15_t energy;         /**< saves previous frame energy. */
+          q15_t x0;             /**< saves previous input sample. */
+  } arm_lms_norm_instance_q15;
+
+
+  /**
+   * @brief Processing function for Q15 normalized LMS filter.
+   * @param[in]  S          points to an instance of the Q15 normalized LMS filter structure.
+   * @param[in]  pSrc       points to the block of input data.
+   * @param[in]  pRef       points to the block of reference data.
+   * @param[out] pOut       points to the block of output data.
+   * @param[out] pErr       points to the block of error data.
+   * @param[in]  blockSize  number of samples to process.
+   */
+  void arm_lms_norm_q15(
+        arm_lms_norm_instance_q15 * S,
+  const q15_t * pSrc,
+        q15_t * pRef,
+        q15_t * pOut,
+        q15_t * pErr,
+        uint32_t blockSize);
+
+
+  /**
+   * @brief Initialization function for Q15 normalized LMS filter.
+   * @param[in] S          points to an instance of the Q15 normalized LMS filter structure.
+   * @param[in] numTaps    number of filter coefficients.
+   * @param[in] pCoeffs    points to coefficient buffer.
+   * @param[in] pState     points to state buffer.
+   * @param[in] mu         step size that controls filter coefficient updates.
+   * @param[in] blockSize  number of samples to process.
+   * @param[in] postShift  bit shift applied to coefficients.
+   */
+  void arm_lms_norm_init_q15(
+        arm_lms_norm_instance_q15 * S,
+        uint16_t numTaps,
+        q15_t * pCoeffs,
+        q15_t * pState,
+        q15_t mu,
+        uint32_t blockSize,
+        uint8_t postShift);
+
+
+  /**
+   * @brief Correlation of floating-point sequences.
+   * @param[in]  pSrcA    points to the first input sequence.
+   * @param[in]  srcALen  length of the first input sequence.
+   * @param[in]  pSrcB    points to the second input sequence.
+   * @param[in]  srcBLen  length of the second input sequence.
+   * @param[out] pDst     points to the block of output data  Length 2 * max(srcALen, srcBLen) - 1.
+   */
+  void arm_correlate_f32(
+  const float32_t * pSrcA,
+        uint32_t srcALen,
+  const float32_t * pSrcB,
+        uint32_t srcBLen,
+        float32_t * pDst);
+
+
+  /**
+   * @brief Correlation of floating-point sequences.
+   * @param[in]  pSrcA    points to the first input sequence.
+   * @param[in]  srcALen  length of the first input sequence.
+   * @param[in]  pSrcB    points to the second input sequence.
+   * @param[in]  srcBLen  length of the second input sequence.
+   * @param[out] pDst     points to the block of output data  Length 2 * max(srcALen, srcBLen) - 1.
+   */
+  void arm_correlate_f64(
+  const float64_t * pSrcA,
+        uint32_t srcALen,
+  const float64_t * pSrcB,
+        uint32_t srcBLen,
+        float64_t * pDst);
+
+
+/**
+ @brief Correlation of Q15 sequences
+ @param[in]  pSrcA     points to the first input sequence
+ @param[in]  srcALen   length of the first input sequence
+ @param[in]  pSrcB     points to the second input sequence
+ @param[in]  srcBLen   length of the second input sequence
+ @param[out] pDst      points to the block of output data  Length 2 * max(srcALen, srcBLen) - 1.
+ @param[in]  pScratch  points to scratch buffer of size max(srcALen, srcBLen) + 2*min(srcALen, srcBLen) - 2.
+*/
+void arm_correlate_opt_q15(
+  const q15_t * pSrcA,
+        uint32_t srcALen,
+  const q15_t * pSrcB,
+        uint32_t srcBLen,
+        q15_t * pDst,
+        q15_t * pScratch);
+
+
+/**
+  @brief Correlation of Q15 sequences.
+  @param[in]  pSrcA    points to the first input sequence
+  @param[in]  srcALen  length of the first input sequence
+  @param[in]  pSrcB    points to the second input sequence
+  @param[in]  srcBLen  length of the second input sequence
+  @param[out] pDst     points to the block of output data  Length 2 * max(srcALen, srcBLen) - 1.
+ */
+  void arm_correlate_q15(
+  const q15_t * pSrcA,
+        uint32_t srcALen,
+  const q15_t * pSrcB,
+        uint32_t srcBLen,
+        q15_t * pDst);
+
+
+/**
+  @brief         Correlation of Q15 sequences (fast version).
+  @param[in]     pSrcA      points to the first input sequence
+  @param[in]     srcALen    length of the first input sequence
+  @param[in]     pSrcB      points to the second input sequence
+  @param[in]     srcBLen    length of the second input sequence
+  @param[out]    pDst       points to the location where the output result is written.  Length 2 * max(srcALen, srcBLen) - 1.
+  @return        none
+ */
+void arm_correlate_fast_q15(
+  const q15_t * pSrcA,
+        uint32_t srcALen,
+  const q15_t * pSrcB,
+        uint32_t srcBLen,
+        q15_t * pDst);
+
+
+/**
+  @brief Correlation of Q15 sequences (fast version).
+  @param[in]  pSrcA     points to the first input sequence.
+  @param[in]  srcALen   length of the first input sequence.
+  @param[in]  pSrcB     points to the second input sequence.
+  @param[in]  srcBLen   length of the second input sequence.
+  @param[out] pDst      points to the block of output data  Length 2 * max(srcALen, srcBLen) - 1.
+  @param[in]  pScratch  points to scratch buffer of size max(srcALen, srcBLen) + 2*min(srcALen, srcBLen) - 2.
+ */
+void arm_correlate_fast_opt_q15(
+  const q15_t * pSrcA,
+        uint32_t srcALen,
+  const q15_t * pSrcB,
+        uint32_t srcBLen,
+        q15_t * pDst,
+        q15_t * pScratch);
+
+
+  /**
+   * @brief Correlation of Q31 sequences.
+   * @param[in]  pSrcA    points to the first input sequence.
+   * @param[in]  srcALen  length of the first input sequence.
+   * @param[in]  pSrcB    points to the second input sequence.
+   * @param[in]  srcBLen  length of the second input sequence.
+   * @param[out] pDst     points to the block of output data  Length 2 * max(srcALen, srcBLen) - 1.
+   */
+  void arm_correlate_q31(
+  const q31_t * pSrcA,
+        uint32_t srcALen,
+  const q31_t * pSrcB,
+        uint32_t srcBLen,
+        q31_t * pDst);
+
+
+/**
+  @brief Correlation of Q31 sequences (fast version).
+  @param[in]  pSrcA    points to the first input sequence
+  @param[in]  srcALen  length of the first input sequence
+  @param[in]  pSrcB    points to the second input sequence
+  @param[in]  srcBLen  length of the second input sequence
+  @param[out] pDst     points to the block of output data  Length 2 * max(srcALen, srcBLen) - 1.
+ */
+void arm_correlate_fast_q31(
+  const q31_t * pSrcA,
+        uint32_t srcALen,
+  const q31_t * pSrcB,
+        uint32_t srcBLen,
+        q31_t * pDst);
+
+
+ /**
+   * @brief Correlation of Q7 sequences.
+   * @param[in]  pSrcA      points to the first input sequence.
+   * @param[in]  srcALen    length of the first input sequence.
+   * @param[in]  pSrcB      points to the second input sequence.
+   * @param[in]  srcBLen    length of the second input sequence.
+   * @param[out] pDst       points to the block of output data  Length 2 * max(srcALen, srcBLen) - 1.
+   * @param[in]  pScratch1  points to scratch buffer(of type q15_t) of size max(srcALen, srcBLen) + 2*min(srcALen, srcBLen) - 2.
+   * @param[in]  pScratch2  points to scratch buffer (of type q15_t) of size min(srcALen, srcBLen).
+   */
+  void arm_correlate_opt_q7(
+  const q7_t * pSrcA,
+        uint32_t srcALen,
+  const q7_t * pSrcB,
+        uint32_t srcBLen,
+        q7_t * pDst,
+        q15_t * pScratch1,
+        q15_t * pScratch2);
+
+
+  /**
+   * @brief Correlation of Q7 sequences.
+   * @param[in]  pSrcA    points to the first input sequence.
+   * @param[in]  srcALen  length of the first input sequence.
+   * @param[in]  pSrcB    points to the second input sequence.
+   * @param[in]  srcBLen  length of the second input sequence.
+   * @param[out] pDst     points to the block of output data  Length 2 * max(srcALen, srcBLen) - 1.
+   */
+  void arm_correlate_q7(
+  const q7_t * pSrcA,
+        uint32_t srcALen,
+  const q7_t * pSrcB,
+        uint32_t srcBLen,
+        q7_t * pDst);
+
+
+  /**
+   * @brief Instance structure for the floating-point sparse FIR filter.
+   */
+  typedef struct
+  {
+          uint16_t numTaps;             /**< number of coefficients in the filter. */
+          uint16_t stateIndex;          /**< state buffer index.  Points to the oldest sample in the state buffer. */
+          float32_t *pState;            /**< points to the state buffer array. The array is of length maxDelay+blockSize-1. */
+    const float32_t *pCoeffs;           /**< points to the coefficient array. The array is of length numTaps.*/
+          uint16_t maxDelay;            /**< maximum offset specified by the pTapDelay array. */
+          int32_t *pTapDelay;           /**< points to the array of delay values.  The array is of length numTaps. */
+  } arm_fir_sparse_instance_f32;
+
+  /**
+   * @brief Instance structure for the Q31 sparse FIR filter.
+   */
+  typedef struct
+  {
+          uint16_t numTaps;             /**< number of coefficients in the filter. */
+          uint16_t stateIndex;          /**< state buffer index.  Points to the oldest sample in the state buffer. */
+          q31_t *pState;                /**< points to the state buffer array. The array is of length maxDelay+blockSize-1. */
+    const q31_t *pCoeffs;               /**< points to the coefficient array. The array is of length numTaps.*/
+          uint16_t maxDelay;            /**< maximum offset specified by the pTapDelay array. */
+          int32_t *pTapDelay;           /**< points to the array of delay values.  The array is of length numTaps. */
+  } arm_fir_sparse_instance_q31;
+
+  /**
+   * @brief Instance structure for the Q15 sparse FIR filter.
+   */
+  typedef struct
+  {
+          uint16_t numTaps;             /**< number of coefficients in the filter. */
+          uint16_t stateIndex;          /**< state buffer index.  Points to the oldest sample in the state buffer. */
+          q15_t *pState;                /**< points to the state buffer array. The array is of length maxDelay+blockSize-1. */
+    const q15_t *pCoeffs;               /**< points to the coefficient array. The array is of length numTaps.*/
+          uint16_t maxDelay;            /**< maximum offset specified by the pTapDelay array. */
+          int32_t *pTapDelay;           /**< points to the array of delay values.  The array is of length numTaps. */
+  } arm_fir_sparse_instance_q15;
+
+  /**
+   * @brief Instance structure for the Q7 sparse FIR filter.
+   */
+  typedef struct
+  {
+          uint16_t numTaps;             /**< number of coefficients in the filter. */
+          uint16_t stateIndex;          /**< state buffer index.  Points to the oldest sample in the state buffer. */
+          q7_t *pState;                 /**< points to the state buffer array. The array is of length maxDelay+blockSize-1. */
+    const q7_t *pCoeffs;                /**< points to the coefficient array. The array is of length numTaps.*/
+          uint16_t maxDelay;            /**< maximum offset specified by the pTapDelay array. */
+          int32_t *pTapDelay;           /**< points to the array of delay values.  The array is of length numTaps. */
+  } arm_fir_sparse_instance_q7;
+
+
+  /**
+   * @brief Processing function for the floating-point sparse FIR filter.
+   * @param[in]  S           points to an instance of the floating-point sparse FIR structure.
+   * @param[in]  pSrc        points to the block of input data.
+   * @param[out] pDst        points to the block of output data
+   * @param[in]  pScratchIn  points to a temporary buffer of size blockSize.
+   * @param[in]  blockSize   number of input samples to process per call.
+   */
+  void arm_fir_sparse_f32(
+        arm_fir_sparse_instance_f32 * S,
+  const float32_t * pSrc,
+        float32_t * pDst,
+        float32_t * pScratchIn,
+        uint32_t blockSize);
+
+
+  /**
+   * @brief  Initialization function for the floating-point sparse FIR filter.
+   * @param[in,out] S          points to an instance of the floating-point sparse FIR structure.
+   * @param[in]     numTaps    number of nonzero coefficients in the filter.
+   * @param[in]     pCoeffs    points to the array of filter coefficients.
+   * @param[in]     pState     points to the state buffer.
+   * @param[in]     pTapDelay  points to the array of offset times.
+   * @param[in]     maxDelay   maximum offset time supported.
+   * @param[in]     blockSize  number of samples that will be processed per block.
+   */
+  void arm_fir_sparse_init_f32(
+        arm_fir_sparse_instance_f32 * S,
+        uint16_t numTaps,
+  const float32_t * pCoeffs,
+        float32_t * pState,
+        int32_t * pTapDelay,
+        uint16_t maxDelay,
+        uint32_t blockSize);
+
+
+  /**
+   * @brief Processing function for the Q31 sparse FIR filter.
+   * @param[in]  S           points to an instance of the Q31 sparse FIR structure.
+   * @param[in]  pSrc        points to the block of input data.
+   * @param[out] pDst        points to the block of output data
+   * @param[in]  pScratchIn  points to a temporary buffer of size blockSize.
+   * @param[in]  blockSize   number of input samples to process per call.
+   */
+  void arm_fir_sparse_q31(
+        arm_fir_sparse_instance_q31 * S,
+  const q31_t * pSrc,
+        q31_t * pDst,
+        q31_t * pScratchIn,
+        uint32_t blockSize);
+
+
+  /**
+   * @brief  Initialization function for the Q31 sparse FIR filter.
+   * @param[in,out] S          points to an instance of the Q31 sparse FIR structure.
+   * @param[in]     numTaps    number of nonzero coefficients in the filter.
+   * @param[in]     pCoeffs    points to the array of filter coefficients.
+   * @param[in]     pState     points to the state buffer.
+   * @param[in]     pTapDelay  points to the array of offset times.
+   * @param[in]     maxDelay   maximum offset time supported.
+   * @param[in]     blockSize  number of samples that will be processed per block.
+   */
+  void arm_fir_sparse_init_q31(
+        arm_fir_sparse_instance_q31 * S,
+        uint16_t numTaps,
+  const q31_t * pCoeffs,
+        q31_t * pState,
+        int32_t * pTapDelay,
+        uint16_t maxDelay,
+        uint32_t blockSize);
+
+
+  /**
+   * @brief Processing function for the Q15 sparse FIR filter.
+   * @param[in]  S            points to an instance of the Q15 sparse FIR structure.
+   * @param[in]  pSrc         points to the block of input data.
+   * @param[out] pDst         points to the block of output data
+   * @param[in]  pScratchIn   points to a temporary buffer of size blockSize.
+   * @param[in]  pScratchOut  points to a temporary buffer of size blockSize.
+   * @param[in]  blockSize    number of input samples to process per call.
+   */
+  void arm_fir_sparse_q15(
+        arm_fir_sparse_instance_q15 * S,
+  const q15_t * pSrc,
+        q15_t * pDst,
+        q15_t * pScratchIn,
+        q31_t * pScratchOut,
+        uint32_t blockSize);
+
+
+  /**
+   * @brief  Initialization function for the Q15 sparse FIR filter.
+   * @param[in,out] S          points to an instance of the Q15 sparse FIR structure.
+   * @param[in]     numTaps    number of nonzero coefficients in the filter.
+   * @param[in]     pCoeffs    points to the array of filter coefficients.
+   * @param[in]     pState     points to the state buffer.
+   * @param[in]     pTapDelay  points to the array of offset times.
+   * @param[in]     maxDelay   maximum offset time supported.
+   * @param[in]     blockSize  number of samples that will be processed per block.
+   */
+  void arm_fir_sparse_init_q15(
+        arm_fir_sparse_instance_q15 * S,
+        uint16_t numTaps,
+  const q15_t * pCoeffs,
+        q15_t * pState,
+        int32_t * pTapDelay,
+        uint16_t maxDelay,
+        uint32_t blockSize);
+
+
+  /**
+   * @brief Processing function for the Q7 sparse FIR filter.
+   * @param[in]  S            points to an instance of the Q7 sparse FIR structure.
+   * @param[in]  pSrc         points to the block of input data.
+   * @param[out] pDst         points to the block of output data
+   * @param[in]  pScratchIn   points to a temporary buffer of size blockSize.
+   * @param[in]  pScratchOut  points to a temporary buffer of size blockSize.
+   * @param[in]  blockSize    number of input samples to process per call.
+   */
+  void arm_fir_sparse_q7(
+        arm_fir_sparse_instance_q7 * S,
+  const q7_t * pSrc,
+        q7_t * pDst,
+        q7_t * pScratchIn,
+        q31_t * pScratchOut,
+        uint32_t blockSize);
+
+
+  /**
+   * @brief  Initialization function for the Q7 sparse FIR filter.
+   * @param[in,out] S          points to an instance of the Q7 sparse FIR structure.
+   * @param[in]     numTaps    number of nonzero coefficients in the filter.
+   * @param[in]     pCoeffs    points to the array of filter coefficients.
+   * @param[in]     pState     points to the state buffer.
+   * @param[in]     pTapDelay  points to the array of offset times.
+   * @param[in]     maxDelay   maximum offset time supported.
+   * @param[in]     blockSize  number of samples that will be processed per block.
+   */
+  void arm_fir_sparse_init_q7(
+        arm_fir_sparse_instance_q7 * S,
+        uint16_t numTaps,
+  const q7_t * pCoeffs,
+        q7_t * pState,
+        int32_t * pTapDelay,
+        uint16_t maxDelay,
+        uint32_t blockSize);
+
+
+
+
+ 
+
+  /**
+   * @brief floating-point Circular write function.
+   */
+  __STATIC_FORCEINLINE void arm_circularWrite_f32(
+  int32_t * circBuffer,
+  int32_t L,
+  uint16_t * writeOffset,
+  int32_t bufferInc,
+  const int32_t * src,
+  int32_t srcInc,
+  uint32_t blockSize)
+  {
+    uint32_t i = 0U;
+    int32_t wOffset;
+
+    /* Copy the value of Index pointer that points
+     * to the current location where the input samples to be copied */
+    wOffset = *writeOffset;
+
+    /* Loop over the blockSize */
+    i = blockSize;
+
+    while (i > 0U)
+    {
+      /* copy the input sample to the circular buffer */
+      circBuffer[wOffset] = *src;
+
+      /* Update the input pointer */
+      src += srcInc;
+
+      /* Circularly update wOffset.  Watch out for positive and negative value */
+      wOffset += bufferInc;
+      if (wOffset >= L)
+        wOffset -= L;
+
+      /* Decrement the loop counter */
+      i--;
+    }
+
+    /* Update the index pointer */
+    *writeOffset = (uint16_t)wOffset;
+  }
+
+
+
+  /**
+   * @brief floating-point Circular Read function.
+   */
+  __STATIC_FORCEINLINE void arm_circularRead_f32(
+  int32_t * circBuffer,
+  int32_t L,
+  int32_t * readOffset,
+  int32_t bufferInc,
+  int32_t * dst,
+  int32_t * dst_base,
+  int32_t dst_length,
+  int32_t dstInc,
+  uint32_t blockSize)
+  {
+    uint32_t i = 0U;
+    int32_t rOffset;
+    int32_t* dst_end;
+
+    /* Copy the value of Index pointer that points
+     * to the current location from where the input samples to be read */
+    rOffset = *readOffset;
+    dst_end = dst_base + dst_length;
+
+    /* Loop over the blockSize */
+    i = blockSize;
+
+    while (i > 0U)
+    {
+      /* copy the sample from the circular buffer to the destination buffer */
+      *dst = circBuffer[rOffset];
+
+      /* Update the input pointer */
+      dst += dstInc;
+
+      if (dst == dst_end)
+      {
+        dst = dst_base;
+      }
+
+      /* Circularly update rOffset.  Watch out for positive and negative value  */
+      rOffset += bufferInc;
+
+      if (rOffset >= L)
+      {
+        rOffset -= L;
+      }
+
+      /* Decrement the loop counter */
+      i--;
+    }
+
+    /* Update the index pointer */
+    *readOffset = rOffset;
+  }
+
+
+  /**
+   * @brief Q15 Circular write function.
+   */
+  __STATIC_FORCEINLINE void arm_circularWrite_q15(
+  q15_t * circBuffer,
+  int32_t L,
+  uint16_t * writeOffset,
+  int32_t bufferInc,
+  const q15_t * src,
+  int32_t srcInc,
+  uint32_t blockSize)
+  {
+    uint32_t i = 0U;
+    int32_t wOffset;
+
+    /* Copy the value of Index pointer that points
+     * to the current location where the input samples to be copied */
+    wOffset = *writeOffset;
+
+    /* Loop over the blockSize */
+    i = blockSize;
+
+    while (i > 0U)
+    {
+      /* copy the input sample to the circular buffer */
+      circBuffer[wOffset] = *src;
+
+      /* Update the input pointer */
+      src += srcInc;
+
+      /* Circularly update wOffset.  Watch out for positive and negative value */
+      wOffset += bufferInc;
+      if (wOffset >= L)
+        wOffset -= L;
+
+      /* Decrement the loop counter */
+      i--;
+    }
+
+    /* Update the index pointer */
+    *writeOffset = (uint16_t)wOffset;
+  }
+
+
+  /**
+   * @brief Q15 Circular Read function.
+   */
+  __STATIC_FORCEINLINE void arm_circularRead_q15(
+  q15_t * circBuffer,
+  int32_t L,
+  int32_t * readOffset,
+  int32_t bufferInc,
+  q15_t * dst,
+  q15_t * dst_base,
+  int32_t dst_length,
+  int32_t dstInc,
+  uint32_t blockSize)
+  {
+    uint32_t i = 0;
+    int32_t rOffset;
+    q15_t* dst_end;
+
+    /* Copy the value of Index pointer that points
+     * to the current location from where the input samples to be read */
+    rOffset = *readOffset;
+
+    dst_end = dst_base + dst_length;
+
+    /* Loop over the blockSize */
+    i = blockSize;
+
+    while (i > 0U)
+    {
+      /* copy the sample from the circular buffer to the destination buffer */
+      *dst = circBuffer[rOffset];
+
+      /* Update the input pointer */
+      dst += dstInc;
+
+      if (dst == dst_end)
+      {
+        dst = dst_base;
+      }
+
+      /* Circularly update wOffset.  Watch out for positive and negative value */
+      rOffset += bufferInc;
+
+      if (rOffset >= L)
+      {
+        rOffset -= L;
+      }
+
+      /* Decrement the loop counter */
+      i--;
+    }
+
+    /* Update the index pointer */
+    *readOffset = rOffset;
+  }
+
+
+  /**
+   * @brief Q7 Circular write function.
+   */
+  __STATIC_FORCEINLINE void arm_circularWrite_q7(
+  q7_t * circBuffer,
+  int32_t L,
+  uint16_t * writeOffset,
+  int32_t bufferInc,
+  const q7_t * src,
+  int32_t srcInc,
+  uint32_t blockSize)
+  {
+    uint32_t i = 0U;
+    int32_t wOffset;
+
+    /* Copy the value of Index pointer that points
+     * to the current location where the input samples to be copied */
+    wOffset = *writeOffset;
+
+    /* Loop over the blockSize */
+    i = blockSize;
+
+    while (i > 0U)
+    {
+      /* copy the input sample to the circular buffer */
+      circBuffer[wOffset] = *src;
+
+      /* Update the input pointer */
+      src += srcInc;
+
+      /* Circularly update wOffset.  Watch out for positive and negative value */
+      wOffset += bufferInc;
+      if (wOffset >= L)
+        wOffset -= L;
+
+      /* Decrement the loop counter */
+      i--;
+    }
+
+    /* Update the index pointer */
+    *writeOffset = (uint16_t)wOffset;
+  }
+
+
+  /**
+   * @brief Q7 Circular Read function.
+   */
+  __STATIC_FORCEINLINE void arm_circularRead_q7(
+  q7_t * circBuffer,
+  int32_t L,
+  int32_t * readOffset,
+  int32_t bufferInc,
+  q7_t * dst,
+  q7_t * dst_base,
+  int32_t dst_length,
+  int32_t dstInc,
+  uint32_t blockSize)
+  {
+    uint32_t i = 0;
+    int32_t rOffset;
+    q7_t* dst_end;
+
+    /* Copy the value of Index pointer that points
+     * to the current location from where the input samples to be read */
+    rOffset = *readOffset;
+
+    dst_end = dst_base + dst_length;
+
+    /* Loop over the blockSize */
+    i = blockSize;
+
+    while (i > 0U)
+    {
+      /* copy the sample from the circular buffer to the destination buffer */
+      *dst = circBuffer[rOffset];
+
+      /* Update the input pointer */
+      dst += dstInc;
+
+      if (dst == dst_end)
+      {
+        dst = dst_base;
+      }
+
+      /* Circularly update rOffset.  Watch out for positive and negative value */
+      rOffset += bufferInc;
+
+      if (rOffset >= L)
+      {
+        rOffset -= L;
+      }
+
+      /* Decrement the loop counter */
+      i--;
+    }
+
+    /* Update the index pointer */
+    *readOffset = rOffset;
+  }
+
+
+/**
+  @brief         Levinson Durbin
+  @param[in]     phi      autocovariance vector starting with lag 0 (length is nbCoefs + 1)
+  @param[out]    a        autoregressive coefficients
+  @param[out]    err      prediction error (variance)
+  @param[in]     nbCoefs  number of autoregressive coefficients
+  @return        none
+ */
+void arm_levinson_durbin_f32(const float32_t *phi,
+  float32_t *a, 
+  float32_t *err,
+  int nbCoefs);
+
+
+/**
+  @brief         Levinson Durbin
+  @param[in]     phi      autocovariance vector starting with lag 0 (length is nbCoefs + 1)
+  @param[out]    a        autoregressive coefficients
+  @param[out]    err      prediction error (variance)
+  @param[in]     nbCoefs  number of autoregressive coefficients
+  @return        none
+ */
+void arm_levinson_durbin_q31(const q31_t *phi,
+  q31_t *a, 
+  q31_t *err,
+  int nbCoefs);
+
+#ifdef   __cplusplus
+}
+#endif
+
+#endif /* ifndef _FILTERING_FUNCTIONS_H_ */

bsp/cmsis/include/dsp/filtering_functions_f16.h

@@ -0,0 +1,237 @@
+/******************************************************************************
+ * @file     filtering_functions_f16.h
+ * @brief    Public header file for CMSIS DSP Library
+ * @version  V1.10.0
+ * @date     08 July 2021
+ * Target Processor: Cortex-M and Cortex-A cores
+ ******************************************************************************/
+/*
+ * Copyright (c) 2010-2020 Arm Limited or its affiliates. All rights reserved.
+ *
+ * SPDX-License-Identifier: Apache-2.0
+ *
+ * Licensed under the Apache License, Version 2.0 (the License); you may
+ * not use this file except in compliance with the License.
+ * You may obtain a copy of the License at
+ *
+ * www.apache.org/licenses/LICENSE-2.0
+ *
+ * Unless required by applicable law or agreed to in writing, software
+ * distributed under the License is distributed on an AS IS BASIS, WITHOUT
+ * WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ * See the License for the specific language governing permissions and
+ * limitations under the License.
+ */
+
+ 
+#ifndef _FILTERING_FUNCTIONS_F16_H_
+#define _FILTERING_FUNCTIONS_F16_H_
+
+#include "arm_math_types_f16.h"
+#include "arm_math_memory.h"
+
+#include "dsp/none.h"
+#include "dsp/utils.h"
+
+
+#ifdef   __cplusplus
+extern "C"
+{
+#endif
+
+#if defined(ARM_FLOAT16_SUPPORTED)
+
+ /**
+   * @brief Instance structure for the floating-point FIR filter.
+   */
+  typedef struct
+  {
+          uint16_t numTaps;     /**< number of filter coefficients in the filter. */
+          float16_t *pState;    /**< points to the state variable array. The array is of length numTaps+blockSize-1. */
+    const float16_t *pCoeffs;   /**< points to the coefficient array. The array is of length numTaps. */
+  } arm_fir_instance_f16;
+
+  /**
+   * @brief  Initialization function for the floating-point FIR filter.
+   * @param[in,out] S          points to an instance of the floating-point FIR filter structure.
+   * @param[in]     numTaps    Number of filter coefficients in the filter.
+   * @param[in]     pCoeffs    points to the filter coefficients.
+   * @param[in]     pState     points to the state buffer.
+   * @param[in]     blockSize  number of samples that are processed at a time.
+   */
+  void arm_fir_init_f16(
+        arm_fir_instance_f16 * S,
+        uint16_t numTaps,
+  const float16_t * pCoeffs,
+        float16_t * pState,
+        uint32_t blockSize);
+
+  /**
+   * @brief Processing function for the floating-point FIR filter.
+   * @param[in]  S          points to an instance of the floating-point FIR structure.
+   * @param[in]  pSrc       points to the block of input data.
+   * @param[out] pDst       points to the block of output data.
+   * @param[in]  blockSize  number of samples to process.
+   */
+  void arm_fir_f16(
+  const arm_fir_instance_f16 * S,
+  const float16_t * pSrc,
+        float16_t * pDst,
+        uint32_t blockSize);
+
+
+  /**
+   * @brief Instance structure for the floating-point Biquad cascade filter.
+   */
+  typedef struct
+  {
+          uint32_t numStages;      /**< number of 2nd order stages in the filter.  Overall order is 2*numStages. */
+          float16_t *pState;       /**< Points to the array of state coefficients.  The array is of length 4*numStages. */
+    const float16_t *pCoeffs;      /**< Points to the array of coefficients.  The array is of length 5*numStages. */
+  } arm_biquad_casd_df1_inst_f16;
+
+#if defined(ARM_MATH_MVEF) && !defined(ARM_MATH_AUTOVECTORIZE)
+  /**
+   * @brief Instance structure for the modified Biquad coefs required by vectorized code.
+   */
+  typedef struct
+  {
+      float16_t coeffs[12][8]; /**< Points to the array of modified coefficients.  The array is of length 32. There is one per stage */
+  } arm_biquad_mod_coef_f16;
+#endif 
+
+  /**
+   * @brief Processing function for the floating-point Biquad cascade filter.
+   * @param[in]  S          points to an instance of the floating-point Biquad cascade structure.
+   * @param[in]  pSrc       points to the block of input data.
+   * @param[out] pDst       points to the block of output data.
+   * @param[in]  blockSize  number of samples to process.
+   */
+  void arm_biquad_cascade_df1_f16(
+  const arm_biquad_casd_df1_inst_f16 * S,
+  const float16_t * pSrc,
+        float16_t * pDst,
+        uint32_t blockSize);
+
+#if defined(ARM_MATH_MVEF) && !defined(ARM_MATH_AUTOVECTORIZE)
+  void arm_biquad_cascade_df1_mve_init_f16(
+      arm_biquad_casd_df1_inst_f16 * S,
+      uint8_t numStages,
+      const float16_t * pCoeffs, 
+      arm_biquad_mod_coef_f16 * pCoeffsMod, 
+      float16_t * pState);
+#endif
+
+  void arm_biquad_cascade_df1_init_f16(
+        arm_biquad_casd_df1_inst_f16 * S,
+        uint8_t numStages,
+  const float16_t * pCoeffs,
+        float16_t * pState);
+
+  /**
+   * @brief Instance structure for the floating-point transposed direct form II Biquad cascade filter.
+   */
+  typedef struct
+  {
+          uint8_t numStages;         /**< number of 2nd order stages in the filter.  Overall order is 2*numStages. */
+          float16_t *pState;         /**< points to the array of state coefficients.  The array is of length 2*numStages. */
+    const float16_t *pCoeffs;        /**< points to the array of coefficients.  The array is of length 5*numStages. */
+  } arm_biquad_cascade_df2T_instance_f16;
+
+  /**
+   * @brief Instance structure for the floating-point transposed direct form II Biquad cascade filter.
+   */
+  typedef struct
+  {
+          uint8_t numStages;         /**< number of 2nd order stages in the filter.  Overall order is 2*numStages. */
+          float16_t *pState;         /**< points to the array of state coefficients.  The array is of length 4*numStages. */
+    const float16_t *pCoeffs;        /**< points to the array of coefficients.  The array is of length 5*numStages. */
+  } arm_biquad_cascade_stereo_df2T_instance_f16;
+
+  /**
+   * @brief Processing function for the floating-point transposed direct form II Biquad cascade filter.
+   * @param[in]  S          points to an instance of the filter data structure.
+   * @param[in]  pSrc       points to the block of input data.
+   * @param[out] pDst       points to the block of output data
+   * @param[in]  blockSize  number of samples to process.
+   */
+  void arm_biquad_cascade_df2T_f16(
+  const arm_biquad_cascade_df2T_instance_f16 * S,
+  const float16_t * pSrc,
+        float16_t * pDst,
+        uint32_t blockSize);
+
+  /**
+   * @brief Processing function for the floating-point transposed direct form II Biquad cascade filter. 2 channels
+   * @param[in]  S          points to an instance of the filter data structure.
+   * @param[in]  pSrc       points to the block of input data.
+   * @param[out] pDst       points to the block of output data
+   * @param[in]  blockSize  number of samples to process.
+   */
+  void arm_biquad_cascade_stereo_df2T_f16(
+  const arm_biquad_cascade_stereo_df2T_instance_f16 * S,
+  const float16_t * pSrc,
+        float16_t * pDst,
+        uint32_t blockSize);
+
+  /**
+   * @brief  Initialization function for the floating-point transposed direct form II Biquad cascade filter.
+   * @param[in,out] S          points to an instance of the filter data structure.
+   * @param[in]     numStages  number of 2nd order stages in the filter.
+   * @param[in]     pCoeffs    points to the filter coefficients.
+   * @param[in]     pState     points to the state buffer.
+   */
+  void arm_biquad_cascade_df2T_init_f16(
+        arm_biquad_cascade_df2T_instance_f16 * S,
+        uint8_t numStages,
+  const float16_t * pCoeffs,
+        float16_t * pState);
+
+  /**
+   * @brief  Initialization function for the floating-point transposed direct form II Biquad cascade filter.
+   * @param[in,out] S          points to an instance of the filter data structure.
+   * @param[in]     numStages  number of 2nd order stages in the filter.
+   * @param[in]     pCoeffs    points to the filter coefficients.
+   * @param[in]     pState     points to the state buffer.
+   */
+  void arm_biquad_cascade_stereo_df2T_init_f16(
+        arm_biquad_cascade_stereo_df2T_instance_f16 * S,
+        uint8_t numStages,
+  const float16_t * pCoeffs,
+        float16_t * pState);
+
+  /**
+   * @brief Correlation of floating-point sequences.
+   * @param[in]  pSrcA    points to the first input sequence.
+   * @param[in]  srcALen  length of the first input sequence.
+   * @param[in]  pSrcB    points to the second input sequence.
+   * @param[in]  srcBLen  length of the second input sequence.
+   * @param[out] pDst     points to the block of output data  Length 2 * max(srcALen, srcBLen) - 1.
+   */
+  void arm_correlate_f16(
+  const float16_t * pSrcA,
+        uint32_t srcALen,
+  const float16_t * pSrcB,
+        uint32_t srcBLen,
+        float16_t * pDst);
+
+
+/**
+  @brief         Levinson Durbin
+  @param[in]     phi      autocovariance vector starting with lag 0 (length is nbCoefs + 1)
+  @param[out]    a        autoregressive coefficients
+  @param[out]    err      prediction error (variance)
+  @param[in]     nbCoefs  number of autoregressive coefficients
+  @return        none
+ */
+void arm_levinson_durbin_f16(const float16_t *phi,
+  float16_t *a, 
+  float16_t *err,
+  int nbCoefs);
+
+#endif /*defined(ARM_FLOAT16_SUPPORTED)*/
+#ifdef   __cplusplus
+}
+#endif
+
+#endif /* ifndef _FILTERING_FUNCTIONS_F16_H_ */

bsp/cmsis/include/dsp/interpolation_functions.h

@@ -0,0 +1,319 @@
+/******************************************************************************
+ * @file     interpolation_functions.h
+ * @brief    Public header file for CMSIS DSP Library
+ * @version  V1.10.0
+ * @date     08 July 2021
+ * Target Processor: Cortex-M and Cortex-A cores
+ ******************************************************************************/
+/*
+ * Copyright (c) 2010-2020 Arm Limited or its affiliates. All rights reserved.
+ *
+ * SPDX-License-Identifier: Apache-2.0
+ *
+ * Licensed under the Apache License, Version 2.0 (the License); you may
+ * not use this file except in compliance with the License.
+ * You may obtain a copy of the License at
+ *
+ * www.apache.org/licenses/LICENSE-2.0
+ *
+ * Unless required by applicable law or agreed to in writing, software
+ * distributed under the License is distributed on an AS IS BASIS, WITHOUT
+ * WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ * See the License for the specific language governing permissions and
+ * limitations under the License.
+ */
+
+ 
+#ifndef _INTERPOLATION_FUNCTIONS_H_
+#define _INTERPOLATION_FUNCTIONS_H_
+
+#include "arm_math_types.h"
+#include "arm_math_memory.h"
+
+#include "dsp/none.h"
+#include "dsp/utils.h"
+
+#ifdef   __cplusplus
+extern "C"
+{
+#endif
+
+
+/**
+ * @defgroup groupInterpolation Interpolation Functions
+ * These functions perform 1- and 2-dimensional interpolation of data.
+ * Linear interpolation is used for 1-dimensional data and
+ * bilinear interpolation is used for 2-dimensional data.
+ */
+
+
+  /**
+   * @brief Instance structure for the floating-point Linear Interpolate function.
+   */
+  typedef struct
+  {
+          uint32_t nValues;           /**< nValues */
+          float32_t x1;               /**< x1 */
+          float32_t xSpacing;         /**< xSpacing */
+          float32_t *pYData;          /**< pointer to the table of Y values */
+  } arm_linear_interp_instance_f32;
+
+  /**
+   * @brief Instance structure for the floating-point bilinear interpolation function.
+   */
+  typedef struct
+  {
+          uint16_t numRows;   /**< number of rows in the data table. */
+          uint16_t numCols;   /**< number of columns in the data table. */
+          float32_t *pData;   /**< points to the data table. */
+  } arm_bilinear_interp_instance_f32;
+
+   /**
+   * @brief Instance structure for the Q31 bilinear interpolation function.
+   */
+  typedef struct
+  {
+          uint16_t numRows;   /**< number of rows in the data table. */
+          uint16_t numCols;   /**< number of columns in the data table. */
+          q31_t *pData;       /**< points to the data table. */
+  } arm_bilinear_interp_instance_q31;
+
+   /**
+   * @brief Instance structure for the Q15 bilinear interpolation function.
+   */
+  typedef struct
+  {
+          uint16_t numRows;   /**< number of rows in the data table. */
+          uint16_t numCols;   /**< number of columns in the data table. */
+          q15_t *pData;       /**< points to the data table. */
+  } arm_bilinear_interp_instance_q15;
+
+   /**
+   * @brief Instance structure for the Q15 bilinear interpolation function.
+   */
+  typedef struct
+  {
+          uint16_t numRows;   /**< number of rows in the data table. */
+          uint16_t numCols;   /**< number of columns in the data table. */
+          q7_t *pData;        /**< points to the data table. */
+  } arm_bilinear_interp_instance_q7;
+
+
+  /**
+   * @brief Struct for specifying cubic spline type
+   */
+  typedef enum
+  {
+    ARM_SPLINE_NATURAL = 0,           /**< Natural spline */
+    ARM_SPLINE_PARABOLIC_RUNOUT = 1   /**< Parabolic runout spline */
+  } arm_spline_type;
+
+  /**
+   * @brief Instance structure for the floating-point cubic spline interpolation.
+   */
+  typedef struct
+  {
+    arm_spline_type type;      /**< Type (boundary conditions) */
+    const float32_t * x;       /**< x values */
+    const float32_t * y;       /**< y values */
+    uint32_t n_x;              /**< Number of known data points */
+    float32_t * coeffs;        /**< Coefficients buffer (b,c, and d) */
+  } arm_spline_instance_f32;
+
+
+
+
+  /**
+   * @ingroup groupInterpolation
+   */
+
+  /**
+   * @addtogroup SplineInterpolate
+   * @{
+   */
+
+  
+  /**
+   * @brief Processing function for the floating-point cubic spline interpolation.
+   * @param[in]  S          points to an instance of the floating-point spline structure.
+   * @param[in]  xq         points to the x values ot the interpolated data points.
+   * @param[out] pDst       points to the block of output data.
+   * @param[in]  blockSize  number of samples of output data.
+   */
+  void arm_spline_f32(
+        arm_spline_instance_f32 * S, 
+  const float32_t * xq,
+        float32_t * pDst,
+        uint32_t blockSize);
+
+  /**
+   * @brief Initialization function for the floating-point cubic spline interpolation.
+   * @param[in,out] S        points to an instance of the floating-point spline structure.
+   * @param[in]     type     type of cubic spline interpolation (boundary conditions)
+   * @param[in]     x        points to the x values of the known data points.
+   * @param[in]     y        points to the y values of the known data points.
+   * @param[in]     n        number of known data points.
+   * @param[in]     coeffs   coefficients array for b, c, and d
+   * @param[in]     tempBuffer   buffer array for internal computations
+   */
+  void arm_spline_init_f32(
+          arm_spline_instance_f32 * S,
+          arm_spline_type type,
+    const float32_t * x,
+    const float32_t * y,
+          uint32_t n, 
+          float32_t * coeffs,
+          float32_t * tempBuffer);
+
+
+  /**
+   * @} end of SplineInterpolate group
+   */
+
+
+  
+  /**
+   * @addtogroup LinearInterpolate
+   * @{
+   */
+
+    /**
+   * @brief  Process function for the floating-point Linear Interpolation Function.
+   * @param[in,out] S  is an instance of the floating-point Linear Interpolation structure
+   * @param[in]     x  input sample to process
+   * @return y processed output sample.
+   *
+   */
+  float32_t arm_linear_interp_f32(
+  arm_linear_interp_instance_f32 * S,
+  float32_t x);
+
+   /**
+   *
+   * @brief  Process function for the Q31 Linear Interpolation Function.
+   * @param[in] pYData   pointer to Q31 Linear Interpolation table
+   * @param[in] x        input sample to process
+   * @param[in] nValues  number of table values
+   * @return y processed output sample.
+   *
+   * \par
+   * Input sample <code>x</code> is in 12.20 format which contains 12 bits for table index and 20 bits for fractional part.
+   * This function can support maximum of table size 2^12.
+   *
+   */
+  q31_t arm_linear_interp_q31(
+  const q31_t * pYData,
+  q31_t x,
+  uint32_t nValues);
+
+  /**
+   *
+   * @brief  Process function for the Q15 Linear Interpolation Function.
+   * @param[in] pYData   pointer to Q15 Linear Interpolation table
+   * @param[in] x        input sample to process
+   * @param[in] nValues  number of table values
+   * @return y processed output sample.
+   *
+   * \par
+   * Input sample <code>x</code> is in 12.20 format which contains 12 bits for table index and 20 bits for fractional part.
+   * This function can support maximum of table size 2^12.
+   *
+   */
+  q15_t arm_linear_interp_q15(
+  const q15_t * pYData,
+  q31_t x,
+  uint32_t nValues);
+
+  /**
+   *
+   * @brief  Process function for the Q7 Linear Interpolation Function.
+   * @param[in] pYData   pointer to Q7 Linear Interpolation table
+   * @param[in] x        input sample to process
+   * @param[in] nValues  number of table values
+   * @return y processed output sample.
+   *
+   * \par
+   * Input sample <code>x</code> is in 12.20 format which contains 12 bits for table index and 20 bits for fractional part.
+   * This function can support maximum of table size 2^12.
+   */
+q7_t arm_linear_interp_q7(
+  const q7_t * pYData,
+  q31_t x,
+  uint32_t nValues);
+
+  /**
+   * @} end of LinearInterpolate group
+   */
+
+  
+
+
+  /**
+   * @ingroup groupInterpolation
+   */
+
+
+  /**
+   * @addtogroup BilinearInterpolate
+   * @{
+   */
+
+  /**
+  * @brief  Floating-point bilinear interpolation.
+  * @param[in,out] S  points to an instance of the interpolation structure.
+  * @param[in]     X  interpolation coordinate.
+  * @param[in]     Y  interpolation coordinate.
+  * @return out interpolated value.
+  */
+  float32_t arm_bilinear_interp_f32(
+  const arm_bilinear_interp_instance_f32 * S,
+  float32_t X,
+  float32_t Y);
+
+  /**
+  * @brief  Q31 bilinear interpolation.
+  * @param[in,out] S  points to an instance of the interpolation structure.
+  * @param[in]     X  interpolation coordinate in 12.20 format.
+  * @param[in]     Y  interpolation coordinate in 12.20 format.
+  * @return out interpolated value.
+  */
+  q31_t arm_bilinear_interp_q31(
+  arm_bilinear_interp_instance_q31 * S,
+  q31_t X,
+  q31_t Y);
+
+
+  /**
+  * @brief  Q15 bilinear interpolation.
+  * @param[in,out] S  points to an instance of the interpolation structure.
+  * @param[in]     X  interpolation coordinate in 12.20 format.
+  * @param[in]     Y  interpolation coordinate in 12.20 format.
+  * @return out interpolated value.
+  */
+  q15_t arm_bilinear_interp_q15(
+  arm_bilinear_interp_instance_q15 * S,
+  q31_t X,
+  q31_t Y);
+
+  /**
+  * @brief  Q7 bilinear interpolation.
+  * @param[in,out] S  points to an instance of the interpolation structure.
+  * @param[in]     X  interpolation coordinate in 12.20 format.
+  * @param[in]     Y  interpolation coordinate in 12.20 format.
+  * @return out interpolated value.
+  */
+  q7_t arm_bilinear_interp_q7(
+  arm_bilinear_interp_instance_q7 * S,
+  q31_t X,
+  q31_t Y);
+  /**
+   * @} end of BilinearInterpolate group
+   */
+
+
+
+#ifdef   __cplusplus
+}
+#endif
+
+#endif /* ifndef _INTERPOLATION_FUNCTIONS_H_ */

bsp/cmsis/include/dsp/interpolation_functions_f16.h

@@ -0,0 +1,107 @@
+/******************************************************************************
+ * @file     interpolation_functions_f16.h
+ * @brief    Public header file for CMSIS DSP Library
+ * @version  V1.10.0
+ * @date     08 July 2021
+ * Target Processor: Cortex-M and Cortex-A cores
+ ******************************************************************************/
+/*
+ * Copyright (c) 2010-2020 Arm Limited or its affiliates. All rights reserved.
+ *
+ * SPDX-License-Identifier: Apache-2.0
+ *
+ * Licensed under the Apache License, Version 2.0 (the License); you may
+ * not use this file except in compliance with the License.
+ * You may obtain a copy of the License at
+ *
+ * www.apache.org/licenses/LICENSE-2.0
+ *
+ * Unless required by applicable law or agreed to in writing, software
+ * distributed under the License is distributed on an AS IS BASIS, WITHOUT
+ * WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ * See the License for the specific language governing permissions and
+ * limitations under the License.
+ */
+
+ 
+#ifndef _INTERPOLATION_FUNCTIONS_F16_H_
+#define _INTERPOLATION_FUNCTIONS_F16_H_
+
+#include "arm_math_types_f16.h"
+#include "arm_math_memory.h"
+
+#include "dsp/none.h"
+#include "dsp/utils.h"
+
+#ifdef   __cplusplus
+extern "C"
+{
+#endif
+
+#if defined(ARM_FLOAT16_SUPPORTED)
+
+typedef struct
+{
+    uint32_t  nValues;        /**< nValues */
+    float16_t x1;             /**< x1 */
+    float16_t xSpacing;       /**< xSpacing */
+    float16_t *pYData;        /**< pointer to the table of Y values */
+} arm_linear_interp_instance_f16;
+
+/**
+ * @brief Instance structure for the floating-point bilinear interpolation function.
+ */
+typedef struct
+{
+    uint16_t  numRows;/**< number of rows in the data table. */
+    uint16_t  numCols;/**< number of columns in the data table. */
+    float16_t *pData; /**< points to the data table. */
+} arm_bilinear_interp_instance_f16;
+
+  /**
+   * @addtogroup LinearInterpolate
+   * @{
+   */
+
+    /**
+   * @brief  Process function for the floating-point Linear Interpolation Function.
+   * @param[in,out] S  is an instance of the floating-point Linear Interpolation structure
+   * @param[in]     x  input sample to process
+   * @return y processed output sample.
+   *
+   */
+  float16_t arm_linear_interp_f16(
+  arm_linear_interp_instance_f16 * S,
+  float16_t x);
+
+    /**
+   * @} end of LinearInterpolate group
+   */
+
+/**
+   * @addtogroup BilinearInterpolate
+   * @{
+   */
+
+  /**
+  * @brief  Floating-point bilinear interpolation.
+  * @param[in,out] S  points to an instance of the interpolation structure.
+  * @param[in]     X  interpolation coordinate.
+  * @param[in]     Y  interpolation coordinate.
+  * @return out interpolated value.
+  */
+  float16_t arm_bilinear_interp_f16(
+  const arm_bilinear_interp_instance_f16 * S,
+  float16_t X,
+  float16_t Y);
+
+
+  /**
+   * @} end of BilinearInterpolate group
+   */
+#endif /*defined(ARM_FLOAT16_SUPPORTED)*/
+#ifdef   __cplusplus
+}
+#endif
+
+#endif /* ifndef _INTERPOLATION_FUNCTIONS_F16_H_ */

bsp/cmsis/include/dsp/matrix_functions.h

@@ -0,0 +1,757 @@
+/******************************************************************************
+ * @file     matrix_functions.h
+ * @brief    Public header file for CMSIS DSP Library
+ * @version  V1.10.0
+ * @date     08 July 2021
+ * Target Processor: Cortex-M and Cortex-A cores
+ ******************************************************************************/
+/*
+ * Copyright (c) 2010-2020 Arm Limited or its affiliates. All rights reserved.
+ *
+ * SPDX-License-Identifier: Apache-2.0
+ *
+ * Licensed under the Apache License, Version 2.0 (the License); you may
+ * not use this file except in compliance with the License.
+ * You may obtain a copy of the License at
+ *
+ * www.apache.org/licenses/LICENSE-2.0
+ *
+ * Unless required by applicable law or agreed to in writing, software
+ * distributed under the License is distributed on an AS IS BASIS, WITHOUT
+ * WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ * See the License for the specific language governing permissions and
+ * limitations under the License.
+ */
+
+ 
+#ifndef _MATRIX_FUNCTIONS_H_
+#define _MATRIX_FUNCTIONS_H_
+
+#include "arm_math_types.h"
+#include "arm_math_memory.h"
+
+#include "dsp/none.h"
+#include "dsp/utils.h"
+
+#ifdef   __cplusplus
+extern "C"
+{
+#endif
+
+/**
+ * @defgroup groupMatrix Matrix Functions
+ *
+ * This set of functions provides basic matrix math operations.
+ * The functions operate on matrix data structures.  For example,
+ * the type
+ * definition for the floating-point matrix structure is shown
+ * below:
+ * <pre>
+ *     typedef struct
+ *     {
+ *       uint16_t numRows;     // number of rows of the matrix.
+ *       uint16_t numCols;     // number of columns of the matrix.
+ *       float32_t *pData;     // points to the data of the matrix.
+ *     } arm_matrix_instance_f32;
+ * </pre>
+ * There are similar definitions for Q15 and Q31 data types.
+ *
+ * The structure specifies the size of the matrix and then points to
+ * an array of data.  The array is of size <code>numRows X numCols</code>
+ * and the values are arranged in row order.  That is, the
+ * matrix element (i, j) is stored at:
+ * <pre>
+ *     pData[i*numCols + j]
+ * </pre>
+ *
+ * \par Init Functions
+ * There is an associated initialization function for each type of matrix
+ * data structure.
+ * The initialization function sets the values of the internal structure fields.
+ * Refer to \ref arm_mat_init_f32(), \ref arm_mat_init_q31() and \ref arm_mat_init_q15()
+ * for floating-point, Q31 and Q15 types,  respectively.
+ *
+ * \par
+ * Use of the initialization function is optional. However, if initialization function is used
+ * then the instance structure cannot be placed into a const data section.
+ * To place the instance structure in a const data
+ * section, manually initialize the data structure.  For example:
+ * <pre>
+ * <code>arm_matrix_instance_f32 S = {nRows, nColumns, pData};</code>
+ * <code>arm_matrix_instance_q31 S = {nRows, nColumns, pData};</code>
+ * <code>arm_matrix_instance_q15 S = {nRows, nColumns, pData};</code>
+ * </pre>
+ * where <code>nRows</code> specifies the number of rows, <code>nColumns</code>
+ * specifies the number of columns, and <code>pData</code> points to the
+ * data array.
+ *
+ * \par Size Checking
+ * By default all of the matrix functions perform size checking on the input and
+ * output matrices. For example, the matrix addition function verifies that the
+ * two input matrices and the output matrix all have the same number of rows and
+ * columns. If the size check fails the functions return:
+ * <pre>
+ *     ARM_MATH_SIZE_MISMATCH
+ * </pre>
+ * Otherwise the functions return
+ * <pre>
+ *     ARM_MATH_SUCCESS
+ * </pre>
+ * There is some overhead associated with this matrix size checking.
+ * The matrix size checking is enabled via the \#define
+ * <pre>
+ *     ARM_MATH_MATRIX_CHECK
+ * </pre>
+ * within the library project settings.  By default this macro is defined
+ * and size checking is enabled. By changing the project settings and
+ * undefining this macro size checking is eliminated and the functions
+ * run a bit faster. With size checking disabled the functions always
+ * return <code>ARM_MATH_SUCCESS</code>.
+ */
+
+  /**
+   * @brief Instance structure for the floating-point matrix structure.
+   */
+  typedef struct
+  {
+    uint16_t numRows;     /**< number of rows of the matrix.     */
+    uint16_t numCols;     /**< number of columns of the matrix.  */
+    float32_t *pData;     /**< points to the data of the matrix. */
+  } arm_matrix_instance_f32;
+ 
+ /**
+   * @brief Instance structure for the floating-point matrix structure.
+   */
+  typedef struct
+  {
+    uint16_t numRows;     /**< number of rows of the matrix.     */
+    uint16_t numCols;     /**< number of columns of the matrix.  */
+    float64_t *pData;     /**< points to the data of the matrix. */
+  } arm_matrix_instance_f64;
+
+ /**
+   * @brief Instance structure for the Q7 matrix structure.
+   */
+  typedef struct
+  {
+    uint16_t numRows;     /**< number of rows of the matrix.     */
+    uint16_t numCols;     /**< number of columns of the matrix.  */
+    q7_t *pData;         /**< points to the data of the matrix. */
+  } arm_matrix_instance_q7;
+
+  /**
+   * @brief Instance structure for the Q15 matrix structure.
+   */
+  typedef struct
+  {
+    uint16_t numRows;     /**< number of rows of the matrix.     */
+    uint16_t numCols;     /**< number of columns of the matrix.  */
+    q15_t *pData;         /**< points to the data of the matrix. */
+  } arm_matrix_instance_q15;
+
+  /**
+   * @brief Instance structure for the Q31 matrix structure.
+   */
+  typedef struct
+  {
+    uint16_t numRows;     /**< number of rows of the matrix.     */
+    uint16_t numCols;     /**< number of columns of the matrix.  */
+    q31_t *pData;         /**< points to the data of the matrix. */
+  } arm_matrix_instance_q31;
+
+  /**
+   * @brief Floating-point matrix addition.
+   * @param[in]  pSrcA  points to the first input matrix structure
+   * @param[in]  pSrcB  points to the second input matrix structure
+   * @param[out] pDst   points to output matrix structure
+   * @return     The function returns either
+   * <code>ARM_MATH_SIZE_MISMATCH</code> or <code>ARM_MATH_SUCCESS</code> based on the outcome of size checking.
+   */
+arm_status arm_mat_add_f32(
+  const arm_matrix_instance_f32 * pSrcA,
+  const arm_matrix_instance_f32 * pSrcB,
+        arm_matrix_instance_f32 * pDst);
+
+  /**
+   * @brief Q15 matrix addition.
+   * @param[in]   pSrcA  points to the first input matrix structure
+   * @param[in]   pSrcB  points to the second input matrix structure
+   * @param[out]  pDst   points to output matrix structure
+   * @return     The function returns either
+   * <code>ARM_MATH_SIZE_MISMATCH</code> or <code>ARM_MATH_SUCCESS</code> based on the outcome of size checking.
+   */
+arm_status arm_mat_add_q15(
+  const arm_matrix_instance_q15 * pSrcA,
+  const arm_matrix_instance_q15 * pSrcB,
+        arm_matrix_instance_q15 * pDst);
+
+  /**
+   * @brief Q31 matrix addition.
+   * @param[in]  pSrcA  points to the first input matrix structure
+   * @param[in]  pSrcB  points to the second input matrix structure
+   * @param[out] pDst   points to output matrix structure
+   * @return     The function returns either
+   * <code>ARM_MATH_SIZE_MISMATCH</code> or <code>ARM_MATH_SUCCESS</code> based on the outcome of size checking.
+   */
+arm_status arm_mat_add_q31(
+  const arm_matrix_instance_q31 * pSrcA,
+  const arm_matrix_instance_q31 * pSrcB,
+        arm_matrix_instance_q31 * pDst);
+
+  /**
+   * @brief Floating-point, complex, matrix multiplication.
+   * @param[in]  pSrcA  points to the first input matrix structure
+   * @param[in]  pSrcB  points to the second input matrix structure
+   * @param[out] pDst   points to output matrix structure
+   * @return     The function returns either
+   * <code>ARM_MATH_SIZE_MISMATCH</code> or <code>ARM_MATH_SUCCESS</code> based on the outcome of size checking.
+   */
+arm_status arm_mat_cmplx_mult_f32(
+  const arm_matrix_instance_f32 * pSrcA,
+  const arm_matrix_instance_f32 * pSrcB,
+        arm_matrix_instance_f32 * pDst);
+
+  /**
+   * @brief Q15, complex,  matrix multiplication.
+   * @param[in]  pSrcA  points to the first input matrix structure
+   * @param[in]  pSrcB  points to the second input matrix structure
+   * @param[out] pDst   points to output matrix structure
+   * @return     The function returns either
+   * <code>ARM_MATH_SIZE_MISMATCH</code> or <code>ARM_MATH_SUCCESS</code> based on the outcome of size checking.
+   */
+arm_status arm_mat_cmplx_mult_q15(
+  const arm_matrix_instance_q15 * pSrcA,
+  const arm_matrix_instance_q15 * pSrcB,
+        arm_matrix_instance_q15 * pDst,
+        q15_t * pScratch);
+
+  /**
+   * @brief Q31, complex, matrix multiplication.
+   * @param[in]  pSrcA  points to the first input matrix structure
+   * @param[in]  pSrcB  points to the second input matrix structure
+   * @param[out] pDst   points to output matrix structure
+   * @return     The function returns either
+   * <code>ARM_MATH_SIZE_MISMATCH</code> or <code>ARM_MATH_SUCCESS</code> based on the outcome of size checking.
+   */
+arm_status arm_mat_cmplx_mult_q31(
+  const arm_matrix_instance_q31 * pSrcA,
+  const arm_matrix_instance_q31 * pSrcB,
+        arm_matrix_instance_q31 * pDst);
+
+  /**
+   * @brief Floating-point matrix transpose.
+   * @param[in]  pSrc  points to the input matrix
+   * @param[out] pDst  points to the output matrix
+   * @return    The function returns either  <code>ARM_MATH_SIZE_MISMATCH</code>
+   * or <code>ARM_MATH_SUCCESS</code> based on the outcome of size checking.
+   */
+arm_status arm_mat_trans_f32(
+  const arm_matrix_instance_f32 * pSrc,
+        arm_matrix_instance_f32 * pDst);
+
+/**
+   * @brief Floating-point matrix transpose.
+   * @param[in]  pSrc  points to the input matrix
+   * @param[out] pDst  points to the output matrix
+   * @return    The function returns either  <code>ARM_MATH_SIZE_MISMATCH</code>
+   * or <code>ARM_MATH_SUCCESS</code> based on the outcome of size checking.
+   */
+arm_status arm_mat_trans_f64(
+  const arm_matrix_instance_f64 * pSrc,
+        arm_matrix_instance_f64 * pDst);
+
+  /**
+   * @brief Floating-point complex matrix transpose.
+   * @param[in]  pSrc  points to the input matrix
+   * @param[out] pDst  points to the output matrix
+   * @return    The function returns either  <code>ARM_MATH_SIZE_MISMATCH</code>
+   * or <code>ARM_MATH_SUCCESS</code> based on the outcome of size checking.
+   */
+arm_status arm_mat_cmplx_trans_f32(
+  const arm_matrix_instance_f32 * pSrc,
+  arm_matrix_instance_f32 * pDst);
+
+
+  /**
+   * @brief Q15 matrix transpose.
+   * @param[in]  pSrc  points to the input matrix
+   * @param[out] pDst  points to the output matrix
+   * @return    The function returns either  <code>ARM_MATH_SIZE_MISMATCH</code>
+   * or <code>ARM_MATH_SUCCESS</code> based on the outcome of size checking.
+   */
+arm_status arm_mat_trans_q15(
+  const arm_matrix_instance_q15 * pSrc,
+        arm_matrix_instance_q15 * pDst);
+
+  /**
+   * @brief Q15 complex matrix transpose.
+   * @param[in]  pSrc  points to the input matrix
+   * @param[out] pDst  points to the output matrix
+   * @return    The function returns either  <code>ARM_MATH_SIZE_MISMATCH</code>
+   * or <code>ARM_MATH_SUCCESS</code> based on the outcome of size checking.
+   */
+arm_status arm_mat_cmplx_trans_q15(
+  const arm_matrix_instance_q15 * pSrc,
+  arm_matrix_instance_q15 * pDst);
+
+  /**
+   * @brief Q7 matrix transpose.
+   * @param[in]  pSrc  points to the input matrix
+   * @param[out] pDst  points to the output matrix
+   * @return    The function returns either  <code>ARM_MATH_SIZE_MISMATCH</code>
+   * or <code>ARM_MATH_SUCCESS</code> based on the outcome of size checking.
+   */
+arm_status arm_mat_trans_q7(
+  const arm_matrix_instance_q7 * pSrc,
+        arm_matrix_instance_q7 * pDst);
+
+  /**
+   * @brief Q31 matrix transpose.
+   * @param[in]  pSrc  points to the input matrix
+   * @param[out] pDst  points to the output matrix
+   * @return    The function returns either  <code>ARM_MATH_SIZE_MISMATCH</code>
+   * or <code>ARM_MATH_SUCCESS</code> based on the outcome of size checking.
+   */
+arm_status arm_mat_trans_q31(
+  const arm_matrix_instance_q31 * pSrc,
+        arm_matrix_instance_q31 * pDst);
+
+  /**
+   * @brief Q31 complex matrix transpose.
+   * @param[in]  pSrc  points to the input matrix
+   * @param[out] pDst  points to the output matrix
+   * @return    The function returns either  <code>ARM_MATH_SIZE_MISMATCH</code>
+   * or <code>ARM_MATH_SUCCESS</code> based on the outcome of size checking.
+   */
+arm_status arm_mat_cmplx_trans_q31(
+  const arm_matrix_instance_q31 * pSrc,
+  arm_matrix_instance_q31 * pDst);
+
+  /**
+   * @brief Floating-point matrix multiplication
+   * @param[in]  pSrcA  points to the first input matrix structure
+   * @param[in]  pSrcB  points to the second input matrix structure
+   * @param[out] pDst   points to output matrix structure
+   * @return     The function returns either
+   * <code>ARM_MATH_SIZE_MISMATCH</code> or <code>ARM_MATH_SUCCESS</code> based on the outcome of size checking.
+   */
+arm_status arm_mat_mult_f32(
+  const arm_matrix_instance_f32 * pSrcA,
+  const arm_matrix_instance_f32 * pSrcB,
+        arm_matrix_instance_f32 * pDst);
+
+  /**
+   * @brief Floating-point matrix multiplication
+   * @param[in]  pSrcA  points to the first input matrix structure
+   * @param[in]  pSrcB  points to the second input matrix structure
+   * @param[out] pDst   points to output matrix structure
+   * @return     The function returns either
+   * <code>ARM_MATH_SIZE_MISMATCH</code> or <code>ARM_MATH_SUCCESS</code> based on the outcome of size checking.
+   */
+arm_status arm_mat_mult_f64(
+  const arm_matrix_instance_f64 * pSrcA,
+  const arm_matrix_instance_f64 * pSrcB,
+        arm_matrix_instance_f64 * pDst);
+
+  /**
+   * @brief Floating-point matrix and vector multiplication
+   * @param[in]  pSrcMat  points to the input matrix structure
+   * @param[in]  pVec     points to vector
+   * @param[out] pDst     points to output vector
+   */
+void arm_mat_vec_mult_f32(
+  const arm_matrix_instance_f32 *pSrcMat, 
+  const float32_t *pVec, 
+  float32_t *pDst);
+
+  /**
+   * @brief Q7 matrix multiplication
+   * @param[in]  pSrcA   points to the first input matrix structure
+   * @param[in]  pSrcB   points to the second input matrix structure
+   * @param[out] pDst    points to output matrix structure
+   * @param[in]  pState  points to the array for storing intermediate results
+   * @return     The function returns either
+   * <code>ARM_MATH_SIZE_MISMATCH</code> or <code>ARM_MATH_SUCCESS</code> based on the outcome of size checking.
+   */
+arm_status arm_mat_mult_q7(
+  const arm_matrix_instance_q7 * pSrcA,
+  const arm_matrix_instance_q7 * pSrcB,
+        arm_matrix_instance_q7 * pDst,
+        q7_t * pState);
+
+  /**
+   * @brief Q7 matrix and vector multiplication
+   * @param[in]  pSrcMat  points to the input matrix structure
+   * @param[in]  pVec     points to vector
+   * @param[out] pDst     points to output vector
+   */
+void arm_mat_vec_mult_q7(
+  const arm_matrix_instance_q7 *pSrcMat, 
+  const q7_t *pVec, 
+  q7_t *pDst);
+
+  /**
+   * @brief Q15 matrix multiplication
+   * @param[in]  pSrcA   points to the first input matrix structure
+   * @param[in]  pSrcB   points to the second input matrix structure
+   * @param[out] pDst    points to output matrix structure
+   * @param[in]  pState  points to the array for storing intermediate results
+   * @return     The function returns either
+   * <code>ARM_MATH_SIZE_MISMATCH</code> or <code>ARM_MATH_SUCCESS</code> based on the outcome of size checking.
+   */
+arm_status arm_mat_mult_q15(
+  const arm_matrix_instance_q15 * pSrcA,
+  const arm_matrix_instance_q15 * pSrcB,
+        arm_matrix_instance_q15 * pDst,
+        q15_t * pState);
+
+  /**
+   * @brief Q15 matrix and vector multiplication
+   * @param[in]  pSrcMat  points to the input matrix structure
+   * @param[in]  pVec     points to vector
+   * @param[out] pDst     points to output vector
+   */
+void arm_mat_vec_mult_q15(
+  const arm_matrix_instance_q15 *pSrcMat, 
+  const q15_t *pVec, 
+  q15_t *pDst);
+
+  /**
+   * @brief Q15 matrix multiplication (fast variant) for Cortex-M3 and Cortex-M4
+   * @param[in]  pSrcA   points to the first input matrix structure
+   * @param[in]  pSrcB   points to the second input matrix structure
+   * @param[out] pDst    points to output matrix structure
+   * @param[in]  pState  points to the array for storing intermediate results
+   * @return     The function returns either
+   * <code>ARM_MATH_SIZE_MISMATCH</code> or <code>ARM_MATH_SUCCESS</code> based on the outcome of size checking.
+   */
+arm_status arm_mat_mult_fast_q15(
+  const arm_matrix_instance_q15 * pSrcA,
+  const arm_matrix_instance_q15 * pSrcB,
+        arm_matrix_instance_q15 * pDst,
+        q15_t * pState);
+
+  /**
+   * @brief Q31 matrix multiplication
+   * @param[in]  pSrcA  points to the first input matrix structure
+   * @param[in]  pSrcB  points to the second input matrix structure
+   * @param[out] pDst   points to output matrix structure
+   * @return     The function returns either
+   * <code>ARM_MATH_SIZE_MISMATCH</code> or <code>ARM_MATH_SUCCESS</code> based on the outcome of size checking.
+   */
+arm_status arm_mat_mult_q31(
+  const arm_matrix_instance_q31 * pSrcA,
+  const arm_matrix_instance_q31 * pSrcB,
+        arm_matrix_instance_q31 * pDst);
+
+  /**
+   * @brief Q31 matrix multiplication
+   * @param[in]  pSrcA  points to the first input matrix structure
+   * @param[in]  pSrcB  points to the second input matrix structure
+   * @param[out] pDst   points to output matrix structure
+   * @param[in]  pState  points to the array for storing intermediate results
+   * @return     The function returns either
+   * <code>ARM_MATH_SIZE_MISMATCH</code> or <code>ARM_MATH_SUCCESS</code> based on the outcome of size checking.
+   */
+arm_status arm_mat_mult_opt_q31(
+  const arm_matrix_instance_q31 * pSrcA,
+  const arm_matrix_instance_q31 * pSrcB,
+        arm_matrix_instance_q31 * pDst,
+        q31_t *pState);
+
+  /**
+   * @brief Q31 matrix and vector multiplication
+   * @param[in]  pSrcMat  points to the input matrix structure
+   * @param[in]  pVec     points to vector
+   * @param[out] pDst     points to output vector
+   */
+void arm_mat_vec_mult_q31(
+  const arm_matrix_instance_q31 *pSrcMat, 
+  const q31_t *pVec, 
+  q31_t *pDst);
+
+  /**
+   * @brief Q31 matrix multiplication (fast variant) for Cortex-M3 and Cortex-M4
+   * @param[in]  pSrcA  points to the first input matrix structure
+   * @param[in]  pSrcB  points to the second input matrix structure
+   * @param[out] pDst   points to output matrix structure
+   * @return     The function returns either
+   * <code>ARM_MATH_SIZE_MISMATCH</code> or <code>ARM_MATH_SUCCESS</code> based on the outcome of size checking.
+   */
+arm_status arm_mat_mult_fast_q31(
+  const arm_matrix_instance_q31 * pSrcA,
+  const arm_matrix_instance_q31 * pSrcB,
+        arm_matrix_instance_q31 * pDst);
+
+  /**
+   * @brief Floating-point matrix subtraction
+   * @param[in]  pSrcA  points to the first input matrix structure
+   * @param[in]  pSrcB  points to the second input matrix structure
+   * @param[out] pDst   points to output matrix structure
+   * @return     The function returns either
+   * <code>ARM_MATH_SIZE_MISMATCH</code> or <code>ARM_MATH_SUCCESS</code> based on the outcome of size checking.
+   */
+arm_status arm_mat_sub_f32(
+  const arm_matrix_instance_f32 * pSrcA,
+  const arm_matrix_instance_f32 * pSrcB,
+        arm_matrix_instance_f32 * pDst);
+
+  /**
+   * @brief Floating-point matrix subtraction
+   * @param[in]  pSrcA  points to the first input matrix structure
+   * @param[in]  pSrcB  points to the second input matrix structure
+   * @param[out] pDst   points to output matrix structure
+   * @return     The function returns either
+   * <code>ARM_MATH_SIZE_MISMATCH</code> or <code>ARM_MATH_SUCCESS</code> based on the outcome of size checking.
+   */
+arm_status arm_mat_sub_f64(
+  const arm_matrix_instance_f64 * pSrcA,
+  const arm_matrix_instance_f64 * pSrcB,
+        arm_matrix_instance_f64 * pDst);
+
+  /**
+   * @brief Q15 matrix subtraction
+   * @param[in]  pSrcA  points to the first input matrix structure
+   * @param[in]  pSrcB  points to the second input matrix structure
+   * @param[out] pDst   points to output matrix structure
+   * @return     The function returns either
+   * <code>ARM_MATH_SIZE_MISMATCH</code> or <code>ARM_MATH_SUCCESS</code> based on the outcome of size checking.
+   */
+arm_status arm_mat_sub_q15(
+  const arm_matrix_instance_q15 * pSrcA,
+  const arm_matrix_instance_q15 * pSrcB,
+        arm_matrix_instance_q15 * pDst);
+
+  /**
+   * @brief Q31 matrix subtraction
+   * @param[in]  pSrcA  points to the first input matrix structure
+   * @param[in]  pSrcB  points to the second input matrix structure
+   * @param[out] pDst   points to output matrix structure
+   * @return     The function returns either
+   * <code>ARM_MATH_SIZE_MISMATCH</code> or <code>ARM_MATH_SUCCESS</code> based on the outcome of size checking.
+   */
+arm_status arm_mat_sub_q31(
+  const arm_matrix_instance_q31 * pSrcA,
+  const arm_matrix_instance_q31 * pSrcB,
+        arm_matrix_instance_q31 * pDst);
+
+  /**
+   * @brief Floating-point matrix scaling.
+   * @param[in]  pSrc   points to the input matrix
+   * @param[in]  scale  scale factor
+   * @param[out] pDst   points to the output matrix
+   * @return     The function returns either
+   * <code>ARM_MATH_SIZE_MISMATCH</code> or <code>ARM_MATH_SUCCESS</code> based on the outcome of size checking.
+   */
+arm_status arm_mat_scale_f32(
+  const arm_matrix_instance_f32 * pSrc,
+        float32_t scale,
+        arm_matrix_instance_f32 * pDst);
+
+  /**
+   * @brief Q15 matrix scaling.
+   * @param[in]  pSrc        points to input matrix
+   * @param[in]  scaleFract  fractional portion of the scale factor
+   * @param[in]  shift       number of bits to shift the result by
+   * @param[out] pDst        points to output matrix
+   * @return     The function returns either
+   * <code>ARM_MATH_SIZE_MISMATCH</code> or <code>ARM_MATH_SUCCESS</code> based on the outcome of size checking.
+   */
+arm_status arm_mat_scale_q15(
+  const arm_matrix_instance_q15 * pSrc,
+        q15_t scaleFract,
+        int32_t shift,
+        arm_matrix_instance_q15 * pDst);
+
+  /**
+   * @brief Q31 matrix scaling.
+   * @param[in]  pSrc        points to input matrix
+   * @param[in]  scaleFract  fractional portion of the scale factor
+   * @param[in]  shift       number of bits to shift the result by
+   * @param[out] pDst        points to output matrix structure
+   * @return     The function returns either
+   * <code>ARM_MATH_SIZE_MISMATCH</code> or <code>ARM_MATH_SUCCESS</code> based on the outcome of size checking.
+   */
+arm_status arm_mat_scale_q31(
+  const arm_matrix_instance_q31 * pSrc,
+        q31_t scaleFract,
+        int32_t shift,
+        arm_matrix_instance_q31 * pDst);
+
+  /**
+   * @brief  Q31 matrix initialization.
+   * @param[in,out] S         points to an instance of the floating-point matrix structure.
+   * @param[in]     nRows     number of rows in the matrix.
+   * @param[in]     nColumns  number of columns in the matrix.
+   * @param[in]     pData     points to the matrix data array.
+   */
+void arm_mat_init_q31(
+        arm_matrix_instance_q31 * S,
+        uint16_t nRows,
+        uint16_t nColumns,
+        q31_t * pData);
+
+  /**
+   * @brief  Q15 matrix initialization.
+   * @param[in,out] S         points to an instance of the floating-point matrix structure.
+   * @param[in]     nRows     number of rows in the matrix.
+   * @param[in]     nColumns  number of columns in the matrix.
+   * @param[in]     pData     points to the matrix data array.
+   */
+void arm_mat_init_q15(
+        arm_matrix_instance_q15 * S,
+        uint16_t nRows,
+        uint16_t nColumns,
+        q15_t * pData);
+
+  /**
+   * @brief  Floating-point matrix initialization.
+   * @param[in,out] S         points to an instance of the floating-point matrix structure.
+   * @param[in]     nRows     number of rows in the matrix.
+   * @param[in]     nColumns  number of columns in the matrix.
+   * @param[in]     pData     points to the matrix data array.
+   */
+void arm_mat_init_f32(
+        arm_matrix_instance_f32 * S,
+        uint16_t nRows,
+        uint16_t nColumns,
+        float32_t * pData);
+
+
+
+  /**
+   * @brief Floating-point matrix inverse.
+   * @param[in]  src   points to the instance of the input floating-point matrix structure.
+   * @param[out] dst   points to the instance of the output floating-point matrix structure.
+   * @return The function returns ARM_MATH_SIZE_MISMATCH, if the dimensions do not match.
+   * If the input matrix is singular (does not have an inverse), then the algorithm terminates and returns error status ARM_MATH_SINGULAR.
+   */
+  arm_status arm_mat_inverse_f32(
+  const arm_matrix_instance_f32 * src,
+  arm_matrix_instance_f32 * dst);
+
+
+  /**
+   * @brief Floating-point matrix inverse.
+   * @param[in]  src   points to the instance of the input floating-point matrix structure.
+   * @param[out] dst   points to the instance of the output floating-point matrix structure.
+   * @return The function returns ARM_MATH_SIZE_MISMATCH, if the dimensions do not match.
+   * If the input matrix is singular (does not have an inverse), then the algorithm terminates and returns error status ARM_MATH_SINGULAR.
+   */
+  arm_status arm_mat_inverse_f64(
+  const arm_matrix_instance_f64 * src,
+  arm_matrix_instance_f64 * dst);
+
+ /**
+   * @brief Floating-point Cholesky decomposition of Symmetric Positive Definite Matrix.
+   * @param[in]  src   points to the instance of the input floating-point matrix structure.
+   * @param[out] dst   points to the instance of the output floating-point matrix structure.
+   * @return The function returns ARM_MATH_SIZE_MISMATCH, if the dimensions do not match.
+   * If the input matrix does not have a decomposition, then the algorithm terminates and returns error status ARM_MATH_DECOMPOSITION_FAILURE.
+   * If the matrix is ill conditioned or only semi-definite, then it is better using the LDL^t decomposition.
+   * The decomposition is returning a lower triangular matrix.
+   */
+  arm_status arm_mat_cholesky_f64(
+  const arm_matrix_instance_f64 * src,
+  arm_matrix_instance_f64 * dst);
+
+ /**
+   * @brief Floating-point Cholesky decomposition of Symmetric Positive Definite Matrix.
+   * @param[in]  src   points to the instance of the input floating-point matrix structure.
+   * @param[out] dst   points to the instance of the output floating-point matrix structure.
+   * @return The function returns ARM_MATH_SIZE_MISMATCH, if the dimensions do not match.
+   * If the input matrix does not have a decomposition, then the algorithm terminates and returns error status ARM_MATH_DECOMPOSITION_FAILURE.
+   * If the matrix is ill conditioned or only semi-definite, then it is better using the LDL^t decomposition.
+   * The decomposition is returning a lower triangular matrix.
+   */
+  arm_status arm_mat_cholesky_f32(
+  const arm_matrix_instance_f32 * src,
+  arm_matrix_instance_f32 * dst);
+
+  /**
+   * @brief Solve UT . X = A where UT is an upper triangular matrix
+   * @param[in]  ut  The upper triangular matrix
+   * @param[in]  a  The matrix a
+   * @param[out] dst The solution X of UT . X = A
+   * @return The function returns ARM_MATH_SINGULAR, if the system can't be solved.
+  */
+  arm_status arm_mat_solve_upper_triangular_f32(
+  const arm_matrix_instance_f32 * ut,
+  const arm_matrix_instance_f32 * a,
+  arm_matrix_instance_f32 * dst);
+
+ /**
+   * @brief Solve LT . X = A where LT is a lower triangular matrix
+   * @param[in]  lt  The lower triangular matrix
+   * @param[in]  a  The matrix a
+   * @param[out] dst The solution X of LT . X = A
+   * @return The function returns ARM_MATH_SINGULAR, if the system can't be solved.
+   */
+  arm_status arm_mat_solve_lower_triangular_f32(
+  const arm_matrix_instance_f32 * lt,
+  const arm_matrix_instance_f32 * a,
+  arm_matrix_instance_f32 * dst);
+
+
+  /**
+   * @brief Solve UT . X = A where UT is an upper triangular matrix
+   * @param[in]  ut  The upper triangular matrix
+   * @param[in]  a  The matrix a
+   * @param[out] dst The solution X of UT . X = A
+   * @return The function returns ARM_MATH_SINGULAR, if the system can't be solved.
+  */
+  arm_status arm_mat_solve_upper_triangular_f64(
+  const arm_matrix_instance_f64 * ut,
+  const arm_matrix_instance_f64 * a,
+  arm_matrix_instance_f64 * dst);
+
+ /**
+   * @brief Solve LT . X = A where LT is a lower triangular matrix
+   * @param[in]  lt  The lower triangular matrix
+   * @param[in]  a  The matrix a
+   * @param[out] dst The solution X of LT . X = A
+   * @return The function returns ARM_MATH_SINGULAR, if the system can't be solved.
+   */
+  arm_status arm_mat_solve_lower_triangular_f64(
+  const arm_matrix_instance_f64 * lt,
+  const arm_matrix_instance_f64 * a,
+  arm_matrix_instance_f64 * dst);
+
+
+  /**
+   * @brief Floating-point LDL decomposition of Symmetric Positive Semi-Definite Matrix.
+   * @param[in]  src   points to the instance of the input floating-point matrix structure.
+   * @param[out] l   points to the instance of the output floating-point triangular matrix structure.
+   * @param[out] d   points to the instance of the output floating-point diagonal matrix structure.
+   * @param[out] p   points to the instance of the output floating-point permutation vector.
+   * @return The function returns ARM_MATH_SIZE_MISMATCH, if the dimensions do not match.
+   * If the input matrix does not have a decomposition, then the algorithm terminates and returns error status ARM_MATH_DECOMPOSITION_FAILURE.
+   * The decomposition is returning a lower triangular matrix.
+   */
+  arm_status arm_mat_ldlt_f32(
+  const arm_matrix_instance_f32 * src,
+  arm_matrix_instance_f32 * l,
+  arm_matrix_instance_f32 * d,
+  uint16_t * pp);
+
+ /**
+   * @brief Floating-point LDL decomposition of Symmetric Positive Semi-Definite Matrix.
+   * @param[in]  src   points to the instance of the input floating-point matrix structure.
+   * @param[out] l   points to the instance of the output floating-point triangular matrix structure.
+   * @param[out] d   points to the instance of the output floating-point diagonal matrix structure.
+   * @param[out] p   points to the instance of the output floating-point permutation vector.
+   * @return The function returns ARM_MATH_SIZE_MISMATCH, if the dimensions do not match.
+   * If the input matrix does not have a decomposition, then the algorithm terminates and returns error status ARM_MATH_DECOMPOSITION_FAILURE.
+   * The decomposition is returning a lower triangular matrix.
+   */
+  arm_status arm_mat_ldlt_f64(
+  const arm_matrix_instance_f64 * src,
+  arm_matrix_instance_f64 * l,
+  arm_matrix_instance_f64 * d,
+  uint16_t * pp);
+
+#ifdef   __cplusplus
+}
+#endif
+
+#endif /* ifndef _MATRIX_FUNCTIONS_H_ */

bsp/cmsis/include/dsp/matrix_functions_f16.h

@@ -0,0 +1,221 @@
+/******************************************************************************
+ * @file     matrix_functions_f16.h
+ * @brief    Public header file for CMSIS DSP Library
+ * @version  V1.10.0
+ * @date     08 July 2021
+ * Target Processor: Cortex-M and Cortex-A cores
+ ******************************************************************************/
+/*
+ * Copyright (c) 2010-2020 Arm Limited or its affiliates. All rights reserved.
+ *
+ * SPDX-License-Identifier: Apache-2.0
+ *
+ * Licensed under the Apache License, Version 2.0 (the License); you may
+ * not use this file except in compliance with the License.
+ * You may obtain a copy of the License at
+ *
+ * www.apache.org/licenses/LICENSE-2.0
+ *
+ * Unless required by applicable law or agreed to in writing, software
+ * distributed under the License is distributed on an AS IS BASIS, WITHOUT
+ * WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ * See the License for the specific language governing permissions and
+ * limitations under the License.
+ */
+
+ 
+#ifndef _MATRIX_FUNCTIONS_F16_H_
+#define _MATRIX_FUNCTIONS_F16_H_
+
+#ifdef   __cplusplus
+extern "C"
+{
+#endif
+
+
+#include "arm_math_types_f16.h"
+#include "arm_math_memory.h"
+
+#include "dsp/none.h"
+#include "dsp/utils.h"
+    
+#if defined(ARM_FLOAT16_SUPPORTED)
+
+ /**
+   * @brief Instance structure for the floating-point matrix structure.
+   */
+  typedef struct
+  {
+    uint16_t numRows;     /**< number of rows of the matrix.     */
+    uint16_t numCols;     /**< number of columns of the matrix.  */
+    float16_t *pData;     /**< points to the data of the matrix. */
+  } arm_matrix_instance_f16;
+
+ /**
+   * @brief Floating-point matrix addition.
+   * @param[in]  pSrcA  points to the first input matrix structure
+   * @param[in]  pSrcB  points to the second input matrix structure
+   * @param[out] pDst   points to output matrix structure
+   * @return     The function returns either
+   * <code>ARM_MATH_SIZE_MISMATCH</code> or <code>ARM_MATH_SUCCESS</code> based on the outcome of size checking.
+   */
+arm_status arm_mat_add_f16(
+  const arm_matrix_instance_f16 * pSrcA,
+  const arm_matrix_instance_f16 * pSrcB,
+        arm_matrix_instance_f16 * pDst);
+
+  /**
+   * @brief Floating-point, complex, matrix multiplication.
+   * @param[in]  pSrcA  points to the first input matrix structure
+   * @param[in]  pSrcB  points to the second input matrix structure
+   * @param[out] pDst   points to output matrix structure
+   * @return     The function returns either
+   * <code>ARM_MATH_SIZE_MISMATCH</code> or <code>ARM_MATH_SUCCESS</code> based on the outcome of size checking.
+   */
+arm_status arm_mat_cmplx_mult_f16(
+  const arm_matrix_instance_f16 * pSrcA,
+  const arm_matrix_instance_f16 * pSrcB,
+        arm_matrix_instance_f16 * pDst);
+
+  /**
+   * @brief Floating-point matrix transpose.
+   * @param[in]  pSrc  points to the input matrix
+   * @param[out] pDst  points to the output matrix
+   * @return    The function returns either  <code>ARM_MATH_SIZE_MISMATCH</code>
+   * or <code>ARM_MATH_SUCCESS</code> based on the outcome of size checking.
+   */
+arm_status arm_mat_trans_f16(
+  const arm_matrix_instance_f16 * pSrc,
+        arm_matrix_instance_f16 * pDst);
+
+  /**
+   * @brief Floating-point complex matrix transpose.
+   * @param[in]  pSrc  points to the input matrix
+   * @param[out] pDst  points to the output matrix
+   * @return    The function returns either  <code>ARM_MATH_SIZE_MISMATCH</code>
+   * or <code>ARM_MATH_SUCCESS</code> based on the outcome of size checking.
+   */
+arm_status arm_mat_cmplx_trans_f16(
+  const arm_matrix_instance_f16 * pSrc,
+  arm_matrix_instance_f16 * pDst);
+
+  /**
+   * @brief Floating-point matrix multiplication
+   * @param[in]  pSrcA  points to the first input matrix structure
+   * @param[in]  pSrcB  points to the second input matrix structure
+   * @param[out] pDst   points to output matrix structure
+   * @return     The function returns either
+   * <code>ARM_MATH_SIZE_MISMATCH</code> or <code>ARM_MATH_SUCCESS</code> based on the outcome of size checking.
+   */
+arm_status arm_mat_mult_f16(
+  const arm_matrix_instance_f16 * pSrcA,
+  const arm_matrix_instance_f16 * pSrcB,
+        arm_matrix_instance_f16 * pDst);
+  /**
+   * @brief Floating-point matrix and vector multiplication
+   * @param[in]  pSrcMat  points to the input matrix structure
+   * @param[in]  pVec     points to vector
+   * @param[out] pDst     points to output vector
+   */
+void arm_mat_vec_mult_f16(
+  const arm_matrix_instance_f16 *pSrcMat, 
+  const float16_t *pVec, 
+  float16_t *pDst);
+
+  /**
+   * @brief Floating-point matrix subtraction
+   * @param[in]  pSrcA  points to the first input matrix structure
+   * @param[in]  pSrcB  points to the second input matrix structure
+   * @param[out] pDst   points to output matrix structure
+   * @return     The function returns either
+   * <code>ARM_MATH_SIZE_MISMATCH</code> or <code>ARM_MATH_SUCCESS</code> based on the outcome of size checking.
+   */
+arm_status arm_mat_sub_f16(
+  const arm_matrix_instance_f16 * pSrcA,
+  const arm_matrix_instance_f16 * pSrcB,
+        arm_matrix_instance_f16 * pDst);
+
+  /**
+   * @brief Floating-point matrix scaling.
+   * @param[in]  pSrc   points to the input matrix
+   * @param[in]  scale  scale factor
+   * @param[out] pDst   points to the output matrix
+   * @return     The function returns either
+   * <code>ARM_MATH_SIZE_MISMATCH</code> or <code>ARM_MATH_SUCCESS</code> based on the outcome of size checking.
+   */
+arm_status arm_mat_scale_f16(
+  const arm_matrix_instance_f16 * pSrc,
+        float16_t scale,
+        arm_matrix_instance_f16 * pDst);
+
+  /**
+   * @brief  Floating-point matrix initialization.
+   * @param[in,out] S         points to an instance of the floating-point matrix structure.
+   * @param[in]     nRows     number of rows in the matrix.
+   * @param[in]     nColumns  number of columns in the matrix.
+   * @param[in]     pData     points to the matrix data array.
+   */
+void arm_mat_init_f16(
+        arm_matrix_instance_f16 * S,
+        uint16_t nRows,
+        uint16_t nColumns,
+        float16_t * pData);
+
+
+  /**
+   * @brief Floating-point matrix inverse.
+   * @param[in]  src   points to the instance of the input floating-point matrix structure.
+   * @param[out] dst   points to the instance of the output floating-point matrix structure.
+   * @return The function returns ARM_MATH_SIZE_MISMATCH, if the dimensions do not match.
+   * If the input matrix is singular (does not have an inverse), then the algorithm terminates and returns error status ARM_MATH_SINGULAR.
+   */
+  arm_status arm_mat_inverse_f16(
+  const arm_matrix_instance_f16 * src,
+  arm_matrix_instance_f16 * dst);
+
+
+ /**
+   * @brief Floating-point Cholesky decomposition of Symmetric Positive Definite Matrix.
+   * @param[in]  src   points to the instance of the input floating-point matrix structure.
+   * @param[out] dst   points to the instance of the output floating-point matrix structure.
+   * @return The function returns ARM_MATH_SIZE_MISMATCH, if the dimensions do not match.
+   * If the input matrix does not have a decomposition, then the algorithm terminates and returns error status ARM_MATH_DECOMPOSITION_FAILURE.
+   * If the matrix is ill conditioned or only semi-definite, then it is better using the LDL^t decomposition.
+   * The decomposition is returning a lower triangular matrix.
+   */
+  arm_status arm_mat_cholesky_f16(
+  const arm_matrix_instance_f16 * src,
+  arm_matrix_instance_f16 * dst);
+
+ /**
+   * @brief Solve UT . X = A where UT is an upper triangular matrix
+   * @param[in]  ut  The upper triangular matrix
+   * @param[in]  a  The matrix a
+   * @param[out] dst The solution X of UT . X = A
+   * @return The function returns ARM_MATH_SINGULAR, if the system can't be solved.
+  */
+  arm_status arm_mat_solve_upper_triangular_f16(
+  const arm_matrix_instance_f16 * ut,
+  const arm_matrix_instance_f16 * a,
+  arm_matrix_instance_f16 * dst);
+
+ /**
+   * @brief Solve LT . X = A where LT is a lower triangular matrix
+   * @param[in]  lt  The lower triangular matrix
+   * @param[in]  a  The matrix a
+   * @param[out] dst The solution X of LT . X = A
+   * @return The function returns ARM_MATH_SINGULAR, if the system can't be solved.
+   */
+  arm_status arm_mat_solve_lower_triangular_f16(
+  const arm_matrix_instance_f16 * lt,
+  const arm_matrix_instance_f16 * a,
+  arm_matrix_instance_f16 * dst);
+
+
+
+#endif /*defined(ARM_FLOAT16_SUPPORTED)*/
+#ifdef   __cplusplus
+}
+#endif
+
+#endif /* ifndef _MATRIX_FUNCTIONS_F16_H_ */

bsp/cmsis/include/dsp/none.h

@@ -0,0 +1,576 @@
+/******************************************************************************
+ * @file     none.h
+ * @brief    Intrinsincs when no DSP extension available
+ * @version  V1.9.0
+ * @date     20. July 2020
+ ******************************************************************************/
+/*
+ * Copyright (c) 2010-2020 Arm Limited or its affiliates. All rights reserved.
+ *
+ * SPDX-License-Identifier: Apache-2.0
+ *
+ * Licensed under the Apache License, Version 2.0 (the License); you may
+ * not use this file except in compliance with the License.
+ * You may obtain a copy of the License at
+ *
+ * www.apache.org/licenses/LICENSE-2.0
+ *
+ * Unless required by applicable law or agreed to in writing, software
+ * distributed under the License is distributed on an AS IS BASIS, WITHOUT
+ * WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ * See the License for the specific language governing permissions and
+ * limitations under the License.
+ */
+
+/*
+
+Definitions in this file are allowing to reuse some versions of the
+CMSIS-DSP to build on a core (M0 for instance) or a host where
+DSP extension are not available.
+
+Ideally a pure C version should have been used instead.
+But those are not always available or use a restricted set
+of intrinsics.
+
+*/
+ 
+#ifndef _NONE_H_
+#define _NONE_H_
+
+#include "arm_math_types.h"
+
+#ifdef   __cplusplus
+extern "C"
+{
+#endif
+
+ 
+
+/*
+
+Normally those kind of definitions are in a compiler file
+in Core or Core_A.
+
+But for MSVC compiler it is a bit special. The goal is very specific
+to CMSIS-DSP and only to allow the use of this library from other
+systems like Python or Matlab.
+
+MSVC is not going to be used to cross-compile to ARM. So, having a MSVC
+compiler file in Core or Core_A would not make sense.
+
+*/
+#if defined ( _MSC_VER ) || defined(__GNUC_PYTHON__) || defined(__APPLE_CC__)
+    __STATIC_FORCEINLINE uint8_t __CLZ(uint32_t data)
+    {
+      if (data == 0U) { return 32U; }
+
+      uint32_t count = 0U;
+      uint32_t mask = 0x80000000U;
+
+      while ((data & mask) == 0U)
+      {
+        count += 1U;
+        mask = mask >> 1U;
+      }
+      return count;
+    }
+
+  __STATIC_FORCEINLINE int32_t __SSAT(int32_t val, uint32_t sat)
+  {
+    if ((sat >= 1U) && (sat <= 32U))
+    {
+      const int32_t max = (int32_t)((1U << (sat - 1U)) - 1U);
+      const int32_t min = -1 - max ;
+      if (val > max)
+      {
+        return max;
+      }
+      else if (val < min)
+      {
+        return min;
+      }
+    }
+    return val;
+  }
+
+  __STATIC_FORCEINLINE uint32_t __USAT(int32_t val, uint32_t sat)
+  {
+    if (sat <= 31U)
+    {
+      const uint32_t max = ((1U << sat) - 1U);
+      if (val > (int32_t)max)
+      {
+        return max;
+      }
+      else if (val < 0)
+      {
+        return 0U;
+      }
+    }
+    return (uint32_t)val;
+  }
+
+ /**
+  \brief   Rotate Right in unsigned value (32 bit)
+  \details Rotate Right (immediate) provides the value of the contents of a register rotated by a variable number of bits.
+  \param [in]    op1  Value to rotate
+  \param [in]    op2  Number of Bits to rotate
+  \return               Rotated value
+ */
+__STATIC_FORCEINLINE uint32_t __ROR(uint32_t op1, uint32_t op2)
+{
+  op2 %= 32U;
+  if (op2 == 0U)
+  {
+    return op1;
+  }
+  return (op1 >> op2) | (op1 << (32U - op2));
+}
+
+
+#endif
+
+/**
+   * @brief Clips Q63 to Q31 values.
+   */
+  __STATIC_FORCEINLINE q31_t clip_q63_to_q31(
+  q63_t x)
+  {
+    return ((q31_t) (x >> 32) != ((q31_t) x >> 31)) ?
+      ((0x7FFFFFFF ^ ((q31_t) (x >> 63)))) : (q31_t) x;
+  }
+
+  /**
+   * @brief Clips Q63 to Q15 values.
+   */
+  __STATIC_FORCEINLINE q15_t clip_q63_to_q15(
+  q63_t x)
+  {
+    return ((q31_t) (x >> 32) != ((q31_t) x >> 31)) ?
+      ((0x7FFF ^ ((q15_t) (x >> 63)))) : (q15_t) (x >> 15);
+  }
+
+  /**
+   * @brief Clips Q31 to Q7 values.
+   */
+  __STATIC_FORCEINLINE q7_t clip_q31_to_q7(
+  q31_t x)
+  {
+    return ((q31_t) (x >> 24) != ((q31_t) x >> 23)) ?
+      ((0x7F ^ ((q7_t) (x >> 31)))) : (q7_t) x;
+  }
+
+  /**
+   * @brief Clips Q31 to Q15 values.
+   */
+  __STATIC_FORCEINLINE q15_t clip_q31_to_q15(
+  q31_t x)
+  {
+    return ((q31_t) (x >> 16) != ((q31_t) x >> 15)) ?
+      ((0x7FFF ^ ((q15_t) (x >> 31)))) : (q15_t) x;
+  }
+
+  /**
+   * @brief Multiplies 32 X 64 and returns 32 bit result in 2.30 format.
+   */
+  __STATIC_FORCEINLINE q63_t mult32x64(
+  q63_t x,
+  q31_t y)
+  {
+    return ((((q63_t) (x & 0x00000000FFFFFFFF) * y) >> 32) +
+            (((q63_t) (x >> 32)                * y)      )  );
+  }
+
+/* SMMLAR */
+#define multAcc_32x32_keep32_R(a, x, y) \
+    a = (q31_t) (((((q63_t) a) << 32) + ((q63_t) x * y) + 0x80000000LL ) >> 32)
+
+/* SMMLSR */
+#define multSub_32x32_keep32_R(a, x, y) \
+    a = (q31_t) (((((q63_t) a) << 32) - ((q63_t) x * y) + 0x80000000LL ) >> 32)
+
+/* SMMULR */
+#define mult_32x32_keep32_R(a, x, y) \
+    a = (q31_t) (((q63_t) x * y + 0x80000000LL ) >> 32)
+
+/* SMMLA */
+#define multAcc_32x32_keep32(a, x, y) \
+    a += (q31_t) (((q63_t) x * y) >> 32)
+
+/* SMMLS */
+#define multSub_32x32_keep32(a, x, y) \
+    a -= (q31_t) (((q63_t) x * y) >> 32)
+
+/* SMMUL */
+#define mult_32x32_keep32(a, x, y) \
+    a = (q31_t) (((q63_t) x * y ) >> 32)
+
+#ifndef ARM_MATH_DSP
+  /**
+   * @brief definition to pack two 16 bit values.
+   */
+  #define __PKHBT(ARG1, ARG2, ARG3) ( (((int32_t)(ARG1) <<    0) & (int32_t)0x0000FFFF) | \
+                                      (((int32_t)(ARG2) << ARG3) & (int32_t)0xFFFF0000)  )
+  #define __PKHTB(ARG1, ARG2, ARG3) ( (((int32_t)(ARG1) <<    0) & (int32_t)0xFFFF0000) | \
+                                      (((int32_t)(ARG2) >> ARG3) & (int32_t)0x0000FFFF)  )
+#endif
+
+   /**
+   * @brief definition to pack four 8 bit values.
+   */
+#ifndef ARM_MATH_BIG_ENDIAN
+  #define __PACKq7(v0,v1,v2,v3) ( (((int32_t)(v0) <<  0) & (int32_t)0x000000FF) | \
+                                  (((int32_t)(v1) <<  8) & (int32_t)0x0000FF00) | \
+                                  (((int32_t)(v2) << 16) & (int32_t)0x00FF0000) | \
+                                  (((int32_t)(v3) << 24) & (int32_t)0xFF000000)  )
+#else
+  #define __PACKq7(v0,v1,v2,v3) ( (((int32_t)(v3) <<  0) & (int32_t)0x000000FF) | \
+                                  (((int32_t)(v2) <<  8) & (int32_t)0x0000FF00) | \
+                                  (((int32_t)(v1) << 16) & (int32_t)0x00FF0000) | \
+                                  (((int32_t)(v0) << 24) & (int32_t)0xFF000000)  )
+#endif
+
+
+ 
+
+/*
+ * @brief C custom defined intrinsic functions
+ */
+#if !defined (ARM_MATH_DSP)
+
+
+  /*
+   * @brief C custom defined QADD8
+   */
+  __STATIC_FORCEINLINE uint32_t __QADD8(
+  uint32_t x,
+  uint32_t y)
+  {
+    q31_t r, s, t, u;
+
+    r = __SSAT(((((q31_t)x << 24) >> 24) + (((q31_t)y << 24) >> 24)), 8) & (int32_t)0x000000FF;
+    s = __SSAT(((((q31_t)x << 16) >> 24) + (((q31_t)y << 16) >> 24)), 8) & (int32_t)0x000000FF;
+    t = __SSAT(((((q31_t)x <<  8) >> 24) + (((q31_t)y <<  8) >> 24)), 8) & (int32_t)0x000000FF;
+    u = __SSAT(((((q31_t)x      ) >> 24) + (((q31_t)y      ) >> 24)), 8) & (int32_t)0x000000FF;
+
+    return ((uint32_t)((u << 24) | (t << 16) | (s <<  8) | (r      )));
+  }
+
+
+  /*
+   * @brief C custom defined QSUB8
+   */
+  __STATIC_FORCEINLINE uint32_t __QSUB8(
+  uint32_t x,
+  uint32_t y)
+  {
+    q31_t r, s, t, u;
+
+    r = __SSAT(((((q31_t)x << 24) >> 24) - (((q31_t)y << 24) >> 24)), 8) & (int32_t)0x000000FF;
+    s = __SSAT(((((q31_t)x << 16) >> 24) - (((q31_t)y << 16) >> 24)), 8) & (int32_t)0x000000FF;
+    t = __SSAT(((((q31_t)x <<  8) >> 24) - (((q31_t)y <<  8) >> 24)), 8) & (int32_t)0x000000FF;
+    u = __SSAT(((((q31_t)x      ) >> 24) - (((q31_t)y      ) >> 24)), 8) & (int32_t)0x000000FF;
+
+    return ((uint32_t)((u << 24) | (t << 16) | (s <<  8) | (r      )));
+  }
+
+
+  /*
+   * @brief C custom defined QADD16
+   */
+  __STATIC_FORCEINLINE uint32_t __QADD16(
+  uint32_t x,
+  uint32_t y)
+  {
+/*  q31_t r,     s;  without initialisation 'arm_offset_q15 test' fails  but 'intrinsic' tests pass! for armCC */
+    q31_t r = 0, s = 0;
+
+    r = __SSAT(((((q31_t)x << 16) >> 16) + (((q31_t)y << 16) >> 16)), 16) & (int32_t)0x0000FFFF;
+    s = __SSAT(((((q31_t)x      ) >> 16) + (((q31_t)y      ) >> 16)), 16) & (int32_t)0x0000FFFF;
+
+    return ((uint32_t)((s << 16) | (r      )));
+  }
+
+
+  /*
+   * @brief C custom defined SHADD16
+   */
+  __STATIC_FORCEINLINE uint32_t __SHADD16(
+  uint32_t x,
+  uint32_t y)
+  {
+    q31_t r, s;
+
+    r = (((((q31_t)x << 16) >> 16) + (((q31_t)y << 16) >> 16)) >> 1) & (int32_t)0x0000FFFF;
+    s = (((((q31_t)x      ) >> 16) + (((q31_t)y      ) >> 16)) >> 1) & (int32_t)0x0000FFFF;
+
+    return ((uint32_t)((s << 16) | (r      )));
+  }
+
+
+  /*
+   * @brief C custom defined QSUB16
+   */
+  __STATIC_FORCEINLINE uint32_t __QSUB16(
+  uint32_t x,
+  uint32_t y)
+  {
+    q31_t r, s;
+
+    r = __SSAT(((((q31_t)x << 16) >> 16) - (((q31_t)y << 16) >> 16)), 16) & (int32_t)0x0000FFFF;
+    s = __SSAT(((((q31_t)x      ) >> 16) - (((q31_t)y      ) >> 16)), 16) & (int32_t)0x0000FFFF;
+
+    return ((uint32_t)((s << 16) | (r      )));
+  }
+
+
+  /*
+   * @brief C custom defined SHSUB16
+   */
+  __STATIC_FORCEINLINE uint32_t __SHSUB16(
+  uint32_t x,
+  uint32_t y)
+  {
+    q31_t r, s;
+
+    r = (((((q31_t)x << 16) >> 16) - (((q31_t)y << 16) >> 16)) >> 1) & (int32_t)0x0000FFFF;
+    s = (((((q31_t)x      ) >> 16) - (((q31_t)y      ) >> 16)) >> 1) & (int32_t)0x0000FFFF;
+
+    return ((uint32_t)((s << 16) | (r      )));
+  }
+
+
+  /*
+   * @brief C custom defined QASX
+   */
+  __STATIC_FORCEINLINE uint32_t __QASX(
+  uint32_t x,
+  uint32_t y)
+  {
+    q31_t r, s;
+
+    r = __SSAT(((((q31_t)x << 16) >> 16) - (((q31_t)y      ) >> 16)), 16) & (int32_t)0x0000FFFF;
+    s = __SSAT(((((q31_t)x      ) >> 16) + (((q31_t)y << 16) >> 16)), 16) & (int32_t)0x0000FFFF;
+
+    return ((uint32_t)((s << 16) | (r      )));
+  }
+
+
+  /*
+   * @brief C custom defined SHASX
+   */
+  __STATIC_FORCEINLINE uint32_t __SHASX(
+  uint32_t x,
+  uint32_t y)
+  {
+    q31_t r, s;
+
+    r = (((((q31_t)x << 16) >> 16) - (((q31_t)y      ) >> 16)) >> 1) & (int32_t)0x0000FFFF;
+    s = (((((q31_t)x      ) >> 16) + (((q31_t)y << 16) >> 16)) >> 1) & (int32_t)0x0000FFFF;
+
+    return ((uint32_t)((s << 16) | (r      )));
+  }
+
+
+  /*
+   * @brief C custom defined QSAX
+   */
+  __STATIC_FORCEINLINE uint32_t __QSAX(
+  uint32_t x,
+  uint32_t y)
+  {
+    q31_t r, s;
+
+    r = __SSAT(((((q31_t)x << 16) >> 16) + (((q31_t)y      ) >> 16)), 16) & (int32_t)0x0000FFFF;
+    s = __SSAT(((((q31_t)x      ) >> 16) - (((q31_t)y << 16) >> 16)), 16) & (int32_t)0x0000FFFF;
+
+    return ((uint32_t)((s << 16) | (r      )));
+  }
+
+
+  /*
+   * @brief C custom defined SHSAX
+   */
+  __STATIC_FORCEINLINE uint32_t __SHSAX(
+  uint32_t x,
+  uint32_t y)
+  {
+    q31_t r, s;
+
+    r = (((((q31_t)x << 16) >> 16) + (((q31_t)y      ) >> 16)) >> 1) & (int32_t)0x0000FFFF;
+    s = (((((q31_t)x      ) >> 16) - (((q31_t)y << 16) >> 16)) >> 1) & (int32_t)0x0000FFFF;
+
+    return ((uint32_t)((s << 16) | (r      )));
+  }
+
+
+  /*
+   * @brief C custom defined SMUSDX
+   */
+  __STATIC_FORCEINLINE uint32_t __SMUSDX(
+  uint32_t x,
+  uint32_t y)
+  {
+    return ((uint32_t)(((((q31_t)x << 16) >> 16) * (((q31_t)y      ) >> 16)) -
+                       ((((q31_t)x      ) >> 16) * (((q31_t)y << 16) >> 16))   ));
+  }
+
+  /*
+   * @brief C custom defined SMUADX
+   */
+  __STATIC_FORCEINLINE uint32_t __SMUADX(
+  uint32_t x,
+  uint32_t y)
+  {
+    return ((uint32_t)(((((q31_t)x << 16) >> 16) * (((q31_t)y      ) >> 16)) +
+                       ((((q31_t)x      ) >> 16) * (((q31_t)y << 16) >> 16))   ));
+  }
+
+
+  /*
+   * @brief C custom defined QADD
+   */
+  __STATIC_FORCEINLINE int32_t __QADD(
+  int32_t x,
+  int32_t y)
+  {
+    return ((int32_t)(clip_q63_to_q31((q63_t)x + (q31_t)y)));
+  }
+
+
+  /*
+   * @brief C custom defined QSUB
+   */
+  __STATIC_FORCEINLINE int32_t __QSUB(
+  int32_t x,
+  int32_t y)
+  {
+    return ((int32_t)(clip_q63_to_q31((q63_t)x - (q31_t)y)));
+  }
+
+
+  /*
+   * @brief C custom defined SMLAD
+   */
+  __STATIC_FORCEINLINE uint32_t __SMLAD(
+  uint32_t x,
+  uint32_t y,
+  uint32_t sum)
+  {
+    return ((uint32_t)(((((q31_t)x << 16) >> 16) * (((q31_t)y << 16) >> 16)) +
+                       ((((q31_t)x      ) >> 16) * (((q31_t)y      ) >> 16)) +
+                       ( ((q31_t)sum    )                                  )   ));
+  }
+
+
+  /*
+   * @brief C custom defined SMLADX
+   */
+  __STATIC_FORCEINLINE uint32_t __SMLADX(
+  uint32_t x,
+  uint32_t y,
+  uint32_t sum)
+  {
+    return ((uint32_t)(((((q31_t)x << 16) >> 16) * (((q31_t)y      ) >> 16)) +
+                       ((((q31_t)x      ) >> 16) * (((q31_t)y << 16) >> 16)) +
+                       ( ((q31_t)sum    )                                  )   ));
+  }
+
+
+  /*
+   * @brief C custom defined SMLSDX
+   */
+  __STATIC_FORCEINLINE uint32_t __SMLSDX(
+  uint32_t x,
+  uint32_t y,
+  uint32_t sum)
+  {
+    return ((uint32_t)(((((q31_t)x << 16) >> 16) * (((q31_t)y      ) >> 16)) -
+                       ((((q31_t)x      ) >> 16) * (((q31_t)y << 16) >> 16)) +
+                       ( ((q31_t)sum    )                                  )   ));
+  }
+
+
+  /*
+   * @brief C custom defined SMLALD
+   */
+  __STATIC_FORCEINLINE uint64_t __SMLALD(
+  uint32_t x,
+  uint32_t y,
+  uint64_t sum)
+  {
+/*  return (sum + ((q15_t) (x >> 16) * (q15_t) (y >> 16)) + ((q15_t) x * (q15_t) y)); */
+    return ((uint64_t)(((((q31_t)x << 16) >> 16) * (((q31_t)y << 16) >> 16)) +
+                       ((((q31_t)x      ) >> 16) * (((q31_t)y      ) >> 16)) +
+                       ( ((q63_t)sum    )                                  )   ));
+  }
+
+
+  /*
+   * @brief C custom defined SMLALDX
+   */
+  __STATIC_FORCEINLINE uint64_t __SMLALDX(
+  uint32_t x,
+  uint32_t y,
+  uint64_t sum)
+  {
+/*  return (sum + ((q15_t) (x >> 16) * (q15_t) y)) + ((q15_t) x * (q15_t) (y >> 16)); */
+    return ((uint64_t)(((((q31_t)x << 16) >> 16) * (((q31_t)y      ) >> 16)) +
+                       ((((q31_t)x      ) >> 16) * (((q31_t)y << 16) >> 16)) +
+                       ( ((q63_t)sum    )                                  )   ));
+  }
+
+
+  /*
+   * @brief C custom defined SMUAD
+   */
+  __STATIC_FORCEINLINE uint32_t __SMUAD(
+  uint32_t x,
+  uint32_t y)
+  {
+    return ((uint32_t)(((((q31_t)x << 16) >> 16) * (((q31_t)y << 16) >> 16)) +
+                       ((((q31_t)x      ) >> 16) * (((q31_t)y      ) >> 16))   ));
+  }
+
+
+  /*
+   * @brief C custom defined SMUSD
+   */
+  __STATIC_FORCEINLINE uint32_t __SMUSD(
+  uint32_t x,
+  uint32_t y)
+  {
+    return ((uint32_t)(((((q31_t)x << 16) >> 16) * (((q31_t)y << 16) >> 16)) -
+                       ((((q31_t)x      ) >> 16) * (((q31_t)y      ) >> 16))   ));
+  }
+
+
+  /*
+   * @brief C custom defined SXTB16
+   */
+  __STATIC_FORCEINLINE uint32_t __SXTB16(
+  uint32_t x)
+  {
+    return ((uint32_t)(((((q31_t)x << 24) >> 24) & (q31_t)0x0000FFFF) |
+                       ((((q31_t)x <<  8) >>  8) & (q31_t)0xFFFF0000)  ));
+  }
+
+  /*
+   * @brief C custom defined SMMLA
+   */
+  __STATIC_FORCEINLINE int32_t __SMMLA(
+  int32_t x,
+  int32_t y,
+  int32_t sum)
+  {
+    return (sum + (int32_t) (((int64_t) x * y) >> 32));
+  }
+
+#endif /* !defined (ARM_MATH_DSP) */
+
+
+#ifdef   __cplusplus
+}
+#endif
+
+#endif /* ifndef _TRANSFORM_FUNCTIONS_H_ */

bsp/cmsis/include/dsp/quaternion_math_functions.h

@@ -0,0 +1,159 @@
+/******************************************************************************
+ * @file     quaternion_math_functions.h
+ * @brief    Public header file for CMSIS DSP Library
+ * @version  V1.10.0
+ * @date     08 July 2021
+ *
+ * Target Processor: Cortex-M and Cortex-A cores
+ ******************************************************************************/
+/*
+ * Copyright (c) 2010-2021 Arm Limited or its affiliates. All rights reserved.
+ *
+ * SPDX-License-Identifier: Apache-2.0
+ *
+ * Licensed under the Apache License, Version 2.0 (the License); you may
+ * not use this file except in compliance with the License.
+ * You may obtain a copy of the License at
+ *
+ * www.apache.org/licenses/LICENSE-2.0
+ *
+ * Unless required by applicable law or agreed to in writing, software
+ * distributed under the License is distributed on an AS IS BASIS, WITHOUT
+ * WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ * See the License for the specific language governing permissions and
+ * limitations under the License.
+ */
+
+ 
+#ifndef _QUATERNION_MATH_FUNCTIONS_H_
+#define _QUATERNION_MATH_FUNCTIONS_H_
+
+#include "arm_math_types.h"
+#include "arm_math_memory.h"
+
+#include "dsp/none.h"
+#include "dsp/utils.h"
+
+
+#ifdef   __cplusplus
+extern "C"
+{
+#endif
+
+/**
+ * @defgroup groupQuaternionMath Quaternion Math Functions
+ * Functions to operates on quaternions and convert between a
+ * rotation and quaternion representation.
+ */
+
+
+/**
+  @brief         Floating-point quaternion Norm.
+  @param[in]     pInputQuaternions       points to the input vector of quaternions
+  @param[out]    pNorms                  points to the output vector of norms
+  @param[in]     nbQuaternions           number of quaternions in each vector
+  @return        none
+ */
+
+
+
+void arm_quaternion_norm_f32(const float32_t *pInputQuaternions, 
+    float32_t *pNorms,
+    uint32_t nbQuaternions);
+
+
+/**
+  @brief         Floating-point quaternion inverse.
+  @param[in]     pInputQuaternions            points to the input vector of quaternions
+  @param[out]    pInverseQuaternions          points to the output vector of inverse quaternions
+  @param[in]     nbQuaternions                number of quaternions in each vector
+  @return        none
+ */
+
+void arm_quaternion_inverse_f32(const float32_t *pInputQuaternions, 
+    float32_t *pInverseQuaternions, 
+    uint32_t nbQuaternions);
+
+/**
+  @brief         Floating-point quaternion conjugates.
+  @param[in]     pInputQuaternions            points to the input vector of quaternions
+  @param[out]    pConjugateQuaternions        points to the output vector of conjugate quaternions
+  @param[in]     nbQuaternions                number of quaternions in each vector
+  @return        none
+ */
+void arm_quaternion_conjugate_f32(const float32_t *inputQuaternions, 
+    float32_t *pConjugateQuaternions, 
+    uint32_t nbQuaternions);
+
+/**
+  @brief         Floating-point normalization of quaternions.
+  @param[in]     pInputQuaternions            points to the input vector of quaternions
+  @param[out]    pNormalizedQuaternions       points to the output vector of normalized quaternions
+  @param[in]     nbQuaternions                number of quaternions in each vector
+  @return        none
+ */
+void arm_quaternion_normalize_f32(const float32_t *inputQuaternions, 
+    float32_t *pNormalizedQuaternions, 
+    uint32_t nbQuaternions);
+
+
+/**
+  @brief         Floating-point product of two quaternions.
+  @param[in]     qa       First quaternion
+  @param[in]     qb       Second quaternion
+  @param[out]    r        Product of two quaternions
+  @return        none
+ */
+void arm_quaternion_product_single_f32(const float32_t *qa, 
+    const float32_t *qb, 
+    float32_t *r);
+
+/**
+  @brief         Floating-point elementwise product two quaternions.
+  @param[in]     qa                  First array of quaternions
+  @param[in]     qb                  Second array of quaternions
+  @param[out]    r                   Elementwise product of quaternions
+  @param[in]     nbQuaternions       Number of quaternions in the array
+  @return        none
+ */
+void arm_quaternion_product_f32(const float32_t *qa, 
+    const float32_t *qb, 
+    float32_t *r,
+    uint32_t nbQuaternions);
+
+/**
+ * @brief Conversion of quaternion to equivalent rotation matrix.
+ * @param[in]       pInputQuaternions points to an array of normalized quaternions
+ * @param[out]      pOutputRotations points to an array of 3x3 rotations (in row order)
+ * @param[in]       nbQuaternions in the array
+ * @return none.
+ *
+ * <b>Format of rotation matrix</b>
+ * \par
+ * The quaternion a + ib + jc + kd is converted into rotation matrix:
+ *   a^2 + b^2 - c^2 - d^2                 2bc - 2ad                 2bd + 2ac
+ *               2bc + 2ad     a^2 - b^2 + c^2 - d^2                 2cd - 2ab
+ *               2bd - 2ac                 2cd + 2ab     a^2 - b^2 - c^2 + d^2
+ *
+ * Rotation matrix is saved in row order : R00 R01 R02 R10 R11 R12 R20 R21 R22
+ */
+void arm_quaternion2rotation_f32(const float32_t *pInputQuaternions, 
+    float32_t *pOutputRotations, 
+    uint32_t nbQuaternions);
+
+/**
+ * @brief Conversion of a rotation matrix to equivalent quaternion.
+ * @param[in]       pInputRotations points to an array 3x3 rotation matrix (in row order)
+ * @param[out]      pOutputQuaternions points to an array of quaternions
+ * @param[in]       nbQuaternions in the array
+ * @return none.
+*/
+void arm_rotation2quaternion_f32(const float32_t *pInputRotations, 
+    float32_t *pOutputQuaternions,  
+    uint32_t nbQuaternions);
+
+#ifdef   __cplusplus
+}
+#endif
+
+#endif /* ifndef _QUATERNION_MATH_FUNCTIONS_H_ */

bsp/cmsis/include/dsp/statistics_functions.h

@@ -0,0 +1,977 @@
+/******************************************************************************
+ * @file     statistics_functions.h
+ * @brief    Public header file for CMSIS DSP Library
+ * @version  V1.10.0
+ * @date     08 July 2021
+ * Target Processor: Cortex-M and Cortex-A cores
+ ******************************************************************************/
+/*
+ * Copyright (c) 2010-2020 Arm Limited or its affiliates. All rights reserved.
+ *
+ * SPDX-License-Identifier: Apache-2.0
+ *
+ * Licensed under the Apache License, Version 2.0 (the License); you may
+ * not use this file except in compliance with the License.
+ * You may obtain a copy of the License at
+ *
+ * www.apache.org/licenses/LICENSE-2.0
+ *
+ * Unless required by applicable law or agreed to in writing, software
+ * distributed under the License is distributed on an AS IS BASIS, WITHOUT
+ * WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ * See the License for the specific language governing permissions and
+ * limitations under the License.
+ */
+
+ 
+#ifndef _STATISTICS_FUNCTIONS_H_
+#define _STATISTICS_FUNCTIONS_H_
+
+#include "arm_math_types.h"
+#include "arm_math_memory.h"
+
+#include "dsp/none.h"
+#include "dsp/utils.h"
+
+#include "dsp/basic_math_functions.h"
+#include "dsp/fast_math_functions.h"
+
+#ifdef   __cplusplus
+extern "C"
+{
+#endif
+
+
+/**
+ * @defgroup groupStats Statistics Functions
+ */
+
+/**
+ * @brief Computation of the LogSumExp
+ *
+ * In probabilistic computations, the dynamic of the probability values can be very
+ * wide because they come from gaussian functions.
+ * To avoid underflow and overflow issues, the values are represented by their log.
+ * In this representation, multiplying the original exp values is easy : their logs are added.
+ * But adding the original exp values is requiring some special handling and it is the
+ * goal of the LogSumExp function.
+ *
+ * If the values are x1...xn, the function is computing:
+ *
+ * ln(exp(x1) + ... + exp(xn)) and the computation is done in such a way that
+ * rounding issues are minimised.
+ *
+ * The max xm of the values is extracted and the function is computing:
+ * xm + ln(exp(x1 - xm) + ... + exp(xn - xm))
+ *
+ * @param[in]  *in         Pointer to an array of input values.
+ * @param[in]  blockSize   Number of samples in the input array.
+ * @return LogSumExp
+ *
+ */
+
+
+float32_t arm_logsumexp_f32(const float32_t *in, uint32_t blockSize);
+
+/**
+ * @brief Dot product with log arithmetic
+ *
+ * Vectors are containing the log of the samples
+ *
+ * @param[in]       pSrcA points to the first input vector
+ * @param[in]       pSrcB points to the second input vector
+ * @param[in]       blockSize number of samples in each vector
+ * @param[in]       pTmpBuffer temporary buffer of length blockSize
+ * @return The log of the dot product .
+ *
+ */
+
+
+float32_t arm_logsumexp_dot_prod_f32(const float32_t * pSrcA,
+  const float32_t * pSrcB,
+  uint32_t blockSize,
+  float32_t *pTmpBuffer);
+
+/**
+ * @brief Entropy
+ *
+ * @param[in]  pSrcA        Array of input values.
+ * @param[in]  blockSize    Number of samples in the input array.
+ * @return     Entropy      -Sum(p ln p)
+ *
+ */
+
+
+float32_t arm_entropy_f32(const float32_t * pSrcA,uint32_t blockSize);
+
+
+/**
+ * @brief Entropy
+ *
+ * @param[in]  pSrcA        Array of input values.
+ * @param[in]  blockSize    Number of samples in the input array.
+ * @return     Entropy      -Sum(p ln p)
+ *
+ */
+
+
+float64_t arm_entropy_f64(const float64_t * pSrcA, uint32_t blockSize);
+
+
+/**
+ * @brief Kullback-Leibler
+ *
+ * @param[in]  pSrcA         Pointer to an array of input values for probability distribution A.
+ * @param[in]  pSrcB         Pointer to an array of input values for probability distribution B.
+ * @param[in]  blockSize     Number of samples in the input array.
+ * @return Kullback-Leibler  Divergence D(A || B)
+ *
+ */
+float32_t arm_kullback_leibler_f32(const float32_t * pSrcA
+  ,const float32_t * pSrcB
+  ,uint32_t blockSize);
+
+
+/**
+ * @brief Kullback-Leibler
+ *
+ * @param[in]  pSrcA         Pointer to an array of input values for probability distribution A.
+ * @param[in]  pSrcB         Pointer to an array of input values for probability distribution B.
+ * @param[in]  blockSize     Number of samples in the input array.
+ * @return Kullback-Leibler  Divergence D(A || B)
+ *
+ */
+float64_t arm_kullback_leibler_f64(const float64_t * pSrcA, 
+                const float64_t * pSrcB, 
+                uint32_t blockSize);
+
+
+ /**
+   * @brief  Sum of the squares of the elements of a Q31 vector.
+   * @param[in]  pSrc       is input pointer
+   * @param[in]  blockSize  is the number of samples to process
+   * @param[out] pResult    is output value.
+   */
+  void arm_power_q31(
+  const q31_t * pSrc,
+        uint32_t blockSize,
+        q63_t * pResult);
+
+
+  /**
+   * @brief  Sum of the squares of the elements of a floating-point vector.
+   * @param[in]  pSrc       is input pointer
+   * @param[in]  blockSize  is the number of samples to process
+   * @param[out] pResult    is output value.
+   */
+  void arm_power_f32(
+  const float32_t * pSrc,
+        uint32_t blockSize,
+        float32_t * pResult);
+
+
+  /**
+   * @brief  Sum of the squares of the elements of a floating-point vector.
+   * @param[in]  pSrc       is input pointer
+   * @param[in]  blockSize  is the number of samples to process
+   * @param[out] pResult    is output value.
+   */
+  void arm_power_f64(
+  const float64_t * pSrc,
+        uint32_t blockSize,
+        float64_t * pResult);
+
+
+  /**
+   * @brief  Sum of the squares of the elements of a Q15 vector.
+   * @param[in]  pSrc       is input pointer
+   * @param[in]  blockSize  is the number of samples to process
+   * @param[out] pResult    is output value.
+   */
+  void arm_power_q15(
+  const q15_t * pSrc,
+        uint32_t blockSize,
+        q63_t * pResult);
+
+
+  /**
+   * @brief  Sum of the squares of the elements of a Q7 vector.
+   * @param[in]  pSrc       is input pointer
+   * @param[in]  blockSize  is the number of samples to process
+   * @param[out] pResult    is output value.
+   */
+  void arm_power_q7(
+  const q7_t * pSrc,
+        uint32_t blockSize,
+        q31_t * pResult);
+
+
+  /**
+   * @brief  Mean value of a Q7 vector.
+   * @param[in]  pSrc       is input pointer
+   * @param[in]  blockSize  is the number of samples to process
+   * @param[out] pResult    is output value.
+   */
+  void arm_mean_q7(
+  const q7_t * pSrc,
+        uint32_t blockSize,
+        q7_t * pResult);
+
+
+  /**
+   * @brief  Mean value of a Q15 vector.
+   * @param[in]  pSrc       is input pointer
+   * @param[in]  blockSize  is the number of samples to process
+   * @param[out] pResult    is output value.
+   */
+  void arm_mean_q15(
+  const q15_t * pSrc,
+        uint32_t blockSize,
+        q15_t * pResult);
+
+
+  /**
+   * @brief  Mean value of a Q31 vector.
+   * @param[in]  pSrc       is input pointer
+   * @param[in]  blockSize  is the number of samples to process
+   * @param[out] pResult    is output value.
+   */
+  void arm_mean_q31(
+  const q31_t * pSrc,
+        uint32_t blockSize,
+        q31_t * pResult);
+
+
+  /**
+   * @brief  Mean value of a floating-point vector.
+   * @param[in]  pSrc       is input pointer
+   * @param[in]  blockSize  is the number of samples to process
+   * @param[out] pResult    is output value.
+   */
+  void arm_mean_f32(
+  const float32_t * pSrc,
+        uint32_t blockSize,
+        float32_t * pResult);
+
+
+  /**
+   * @brief  Mean value of a floating-point vector.
+   * @param[in]  pSrc       is input pointer
+   * @param[in]  blockSize  is the number of samples to process
+   * @param[out] pResult    is output value.
+   */
+  void arm_mean_f64(
+  const float64_t * pSrc,
+        uint32_t blockSize,
+        float64_t * pResult);
+
+
+  /**
+   * @brief  Variance of the elements of a floating-point vector.
+   * @param[in]  pSrc       is input pointer
+   * @param[in]  blockSize  is the number of samples to process
+   * @param[out] pResult    is output value.
+   */
+  void arm_var_f32(
+  const float32_t * pSrc,
+        uint32_t blockSize,
+        float32_t * pResult);
+
+
+  /**
+   * @brief  Variance of the elements of a floating-point vector.
+   * @param[in]  pSrc       is input pointer
+   * @param[in]  blockSize  is the number of samples to process
+   * @param[out] pResult    is output value.
+   */
+  void arm_var_f64(
+  const float64_t * pSrc,
+        uint32_t blockSize,
+        float64_t * pResult);
+
+
+  /**
+   * @brief  Variance of the elements of a Q31 vector.
+   * @param[in]  pSrc       is input pointer
+   * @param[in]  blockSize  is the number of samples to process
+   * @param[out] pResult    is output value.
+   */
+  void arm_var_q31(
+  const q31_t * pSrc,
+        uint32_t blockSize,
+        q31_t * pResult);
+
+
+  /**
+   * @brief  Variance of the elements of a Q15 vector.
+   * @param[in]  pSrc       is input pointer
+   * @param[in]  blockSize  is the number of samples to process
+   * @param[out] pResult    is output value.
+   */
+  void arm_var_q15(
+  const q15_t * pSrc,
+        uint32_t blockSize,
+        q15_t * pResult);
+
+
+  /**
+   * @brief  Root Mean Square of the elements of a floating-point vector.
+   * @param[in]  pSrc       is input pointer
+   * @param[in]  blockSize  is the number of samples to process
+   * @param[out] pResult    is output value.
+   */
+  void arm_rms_f32(
+  const float32_t * pSrc,
+        uint32_t blockSize,
+        float32_t * pResult);
+
+
+  /**
+   * @brief  Root Mean Square of the elements of a Q31 vector.
+   * @param[in]  pSrc       is input pointer
+   * @param[in]  blockSize  is the number of samples to process
+   * @param[out] pResult    is output value.
+   */
+  void arm_rms_q31(
+  const q31_t * pSrc,
+        uint32_t blockSize,
+        q31_t * pResult);
+
+
+  /**
+   * @brief  Root Mean Square of the elements of a Q15 vector.
+   * @param[in]  pSrc       is input pointer
+   * @param[in]  blockSize  is the number of samples to process
+   * @param[out] pResult    is output value.
+   */
+  void arm_rms_q15(
+  const q15_t * pSrc,
+        uint32_t blockSize,
+        q15_t * pResult);
+
+
+  /**
+   * @brief  Standard deviation of the elements of a floating-point vector.
+   * @param[in]  pSrc       is input pointer
+   * @param[in]  blockSize  is the number of samples to process
+   * @param[out] pResult    is output value.
+   */
+  void arm_std_f32(
+  const float32_t * pSrc,
+        uint32_t blockSize,
+        float32_t * pResult);
+
+
+  /**
+   * @brief  Standard deviation of the elements of a floating-point vector.
+   * @param[in]  pSrc       is input pointer
+   * @param[in]  blockSize  is the number of samples to process
+   * @param[out] pResult    is output value.
+   */
+  void arm_std_f64(
+  const float64_t * pSrc,
+        uint32_t blockSize,
+        float64_t * pResult);
+
+
+  /**
+   * @brief  Standard deviation of the elements of a Q31 vector.
+   * @param[in]  pSrc       is input pointer
+   * @param[in]  blockSize  is the number of samples to process
+   * @param[out] pResult    is output value.
+   */
+  void arm_std_q31(
+  const q31_t * pSrc,
+        uint32_t blockSize,
+        q31_t * pResult);
+
+
+  /**
+   * @brief  Standard deviation of the elements of a Q15 vector.
+   * @param[in]  pSrc       is input pointer
+   * @param[in]  blockSize  is the number of samples to process
+   * @param[out] pResult    is output value.
+   */
+  void arm_std_q15(
+  const q15_t * pSrc,
+        uint32_t blockSize,
+        q15_t * pResult);
+
+
+  
+  /**
+   * @brief  Minimum value of a Q7 vector.
+   * @param[in]  pSrc       is input pointer
+   * @param[in]  blockSize  is the number of samples to process
+   * @param[out] result     is output pointer
+   * @param[in]  index      is the array index of the minimum value in the input buffer.
+   */
+  void arm_min_q7(
+  const q7_t * pSrc,
+        uint32_t blockSize,
+        q7_t * result,
+        uint32_t * index);
+
+  /**
+   * @brief  Minimum value of absolute values of a Q7 vector.
+   * @param[in]  pSrc       is input pointer
+   * @param[in]  blockSize  is the number of samples to process
+   * @param[out] result     is output pointer
+   * @param[in]  index      is the array index of the minimum value in the input buffer.
+   */
+  void arm_absmin_q7(
+  const q7_t * pSrc,
+        uint32_t blockSize,
+        q7_t * result,
+        uint32_t * index);
+
+    /**
+   * @brief  Minimum value of absolute values of a Q7 vector.
+   * @param[in]  pSrc       is input pointer
+   * @param[in]  blockSize  is the number of samples to process
+   * @param[out] result     is output pointer
+   */
+  void arm_absmin_no_idx_q7(
+  const q7_t * pSrc,
+        uint32_t blockSize,
+        q7_t * result);
+
+
+  /**
+   * @brief  Minimum value of a Q15 vector.
+   * @param[in]  pSrc       is input pointer
+   * @param[in]  blockSize  is the number of samples to process
+   * @param[out] pResult    is output pointer
+   * @param[in]  pIndex     is the array index of the minimum value in the input buffer.
+   */
+  void arm_min_q15(
+  const q15_t * pSrc,
+        uint32_t blockSize,
+        q15_t * pResult,
+        uint32_t * pIndex);
+
+/**
+   * @brief  Minimum value of absolute values of a Q15 vector.
+   * @param[in]  pSrc       is input pointer
+   * @param[in]  blockSize  is the number of samples to process
+   * @param[out] pResult    is output pointer
+   * @param[in]  pIndex     is the array index of the minimum value in the input buffer.
+   */
+  void arm_absmin_q15(
+  const q15_t * pSrc,
+        uint32_t blockSize,
+        q15_t * pResult,
+        uint32_t * pIndex);
+
+  /**
+   * @brief  Minimum value of absolute values of a Q15 vector.
+   * @param[in]  pSrc       is input pointer
+   * @param[in]  blockSize  is the number of samples to process
+   * @param[out] pResult    is output pointer
+   */
+  void arm_absmin_no_idx_q15(
+  const q15_t * pSrc,
+        uint32_t blockSize,
+        q15_t * pResult);
+
+
+  /**
+   * @brief  Minimum value of a Q31 vector.
+   * @param[in]  pSrc       is input pointer
+   * @param[in]  blockSize  is the number of samples to process
+   * @param[out] pResult    is output pointer
+   * @param[out] pIndex     is the array index of the minimum value in the input buffer.
+   */
+  void arm_min_q31(
+  const q31_t * pSrc,
+        uint32_t blockSize,
+        q31_t * pResult,
+        uint32_t * pIndex);
+
+  /**
+   * @brief  Minimum value of absolute values of a Q31 vector.
+   * @param[in]  pSrc       is input pointer
+   * @param[in]  blockSize  is the number of samples to process
+   * @param[out] pResult    is output pointer
+   * @param[out] pIndex     is the array index of the minimum value in the input buffer.
+   */
+  void arm_absmin_q31(
+  const q31_t * pSrc,
+        uint32_t blockSize,
+        q31_t * pResult,
+        uint32_t * pIndex);
+
+ /**
+   * @brief  Minimum value of absolute values of a Q31 vector.
+   * @param[in]  pSrc       is input pointer
+   * @param[in]  blockSize  is the number of samples to process
+   * @param[out] pResult    is output pointer
+   */
+  void arm_absmin_no_idx_q31(
+  const q31_t * pSrc,
+        uint32_t blockSize,
+        q31_t * pResult);
+
+
+  /**
+   * @brief  Minimum value of a floating-point vector.
+   * @param[in]  pSrc       is input pointer
+   * @param[in]  blockSize  is the number of samples to process
+   * @param[out] pResult    is output pointer
+   * @param[out] pIndex     is the array index of the minimum value in the input buffer.
+   */
+  void arm_min_f32(
+  const float32_t * pSrc,
+        uint32_t blockSize,
+        float32_t * pResult,
+        uint32_t * pIndex);
+
+  /**
+   * @brief  Minimum value of absolute values of a floating-point vector.
+   * @param[in]  pSrc       is input pointer
+   * @param[in]  blockSize  is the number of samples to process
+   * @param[out] pResult    is output pointer
+   * @param[out] pIndex     is the array index of the minimum value in the input buffer.
+   */
+  void arm_absmin_f32(
+  const float32_t * pSrc,
+        uint32_t blockSize,
+        float32_t * pResult,
+        uint32_t * pIndex);
+
+  /**
+   * @brief  Minimum value of absolute values of a floating-point vector.
+   * @param[in]  pSrc       is input pointer
+   * @param[in]  blockSize  is the number of samples to process
+   * @param[out] pResult    is output pointer
+   */
+  void arm_absmin_no_idx_f32(
+  const float32_t * pSrc,
+        uint32_t blockSize,
+        float32_t * pResult);
+
+
+  /**
+   * @brief  Minimum value of a floating-point vector.
+   * @param[in]  pSrc       is input pointer
+   * @param[in]  blockSize  is the number of samples to process
+   * @param[out] pResult    is output pointer
+   * @param[out] pIndex     is the array index of the minimum value in the input buffer.
+   */
+  void arm_min_f64(
+  const float64_t * pSrc,
+        uint32_t blockSize,
+        float64_t * pResult,
+        uint32_t * pIndex);
+
+  /**
+   * @brief  Minimum value of absolute values of a floating-point vector.
+   * @param[in]  pSrc       is input pointer
+   * @param[in]  blockSize  is the number of samples to process
+   * @param[out] pResult    is output pointer
+   * @param[out] pIndex     is the array index of the minimum value in the input buffer.
+   */
+  void arm_absmin_f64(
+  const float64_t * pSrc,
+        uint32_t blockSize,
+        float64_t * pResult,
+        uint32_t * pIndex);
+
+  /**
+   * @brief  Minimum value of absolute values of a floating-point vector.
+   * @param[in]  pSrc       is input pointer
+   * @param[in]  blockSize  is the number of samples to process
+   * @param[out] pResult    is output pointer
+   */
+  void arm_absmin_no_idx_f64(
+  const float64_t * pSrc,
+        uint32_t blockSize,
+        float64_t * pResult);
+
+
+/**
+ * @brief Maximum value of a Q7 vector.
+ * @param[in]  pSrc       points to the input buffer
+ * @param[in]  blockSize  length of the input vector
+ * @param[out] pResult    maximum value returned here
+ * @param[out] pIndex     index of maximum value returned here
+ */
+  void arm_max_q7(
+  const q7_t * pSrc,
+        uint32_t blockSize,
+        q7_t * pResult,
+        uint32_t * pIndex);
+
+/**
+ * @brief Maximum value of absolute values of a Q7 vector.
+ * @param[in]  pSrc       points to the input buffer
+ * @param[in]  blockSize  length of the input vector
+ * @param[out] pResult    maximum value returned here
+ * @param[out] pIndex     index of maximum value returned here
+ */
+  void arm_absmax_q7(
+  const q7_t * pSrc,
+        uint32_t blockSize,
+        q7_t * pResult,
+        uint32_t * pIndex);
+
+/**
+ * @brief Maximum value of absolute values of a Q7 vector.
+ * @param[in]  pSrc       points to the input buffer
+ * @param[in]  blockSize  length of the input vector
+ * @param[out] pResult    maximum value returned here
+ */
+  void arm_absmax_no_idx_q7(
+  const q7_t * pSrc,
+        uint32_t blockSize,
+        q7_t * pResult);
+
+
+/**
+ * @brief Maximum value of a Q15 vector.
+ * @param[in]  pSrc       points to the input buffer
+ * @param[in]  blockSize  length of the input vector
+ * @param[out] pResult    maximum value returned here
+ * @param[out] pIndex     index of maximum value returned here
+ */
+  void arm_max_q15(
+  const q15_t * pSrc,
+        uint32_t blockSize,
+        q15_t * pResult,
+        uint32_t * pIndex);
+
+/**
+ * @brief Maximum value of absolute values of a Q15 vector.
+ * @param[in]  pSrc       points to the input buffer
+ * @param[in]  blockSize  length of the input vector
+ * @param[out] pResult    maximum value returned here
+ * @param[out] pIndex     index of maximum value returned here
+ */
+  void arm_absmax_q15(
+  const q15_t * pSrc,
+        uint32_t blockSize,
+        q15_t * pResult,
+        uint32_t * pIndex);
+
+  /**
+ * @brief Maximum value of absolute values of a Q15 vector.
+ * @param[in]  pSrc       points to the input buffer
+ * @param[in]  blockSize  length of the input vector
+ * @param[out] pResult    maximum value returned here
+ */
+  void arm_absmax_no_idx_q15(
+  const q15_t * pSrc,
+        uint32_t blockSize,
+        q15_t * pResult);
+
+/**
+ * @brief Maximum value of a Q31 vector.
+ * @param[in]  pSrc       points to the input buffer
+ * @param[in]  blockSize  length of the input vector
+ * @param[out] pResult    maximum value returned here
+ * @param[out] pIndex     index of maximum value returned here
+ */
+  void arm_max_q31(
+  const q31_t * pSrc,
+        uint32_t blockSize,
+        q31_t * pResult,
+        uint32_t * pIndex);
+
+/**
+ * @brief Maximum value of absolute values of a Q31 vector.
+ * @param[in]  pSrc       points to the input buffer
+ * @param[in]  blockSize  length of the input vector
+ * @param[out] pResult    maximum value returned here
+ * @param[out] pIndex     index of maximum value returned here
+ */
+  void arm_absmax_q31(
+  const q31_t * pSrc,
+        uint32_t blockSize,
+        q31_t * pResult,
+        uint32_t * pIndex);
+
+ /**
+ * @brief Maximum value of absolute values of a Q31 vector.
+ * @param[in]  pSrc       points to the input buffer
+ * @param[in]  blockSize  length of the input vector
+ * @param[out] pResult    maximum value returned here
+ */
+  void arm_absmax_no_idx_q31(
+  const q31_t * pSrc,
+        uint32_t blockSize,
+        q31_t * pResult);
+
+/**
+ * @brief Maximum value of a floating-point vector.
+ * @param[in]  pSrc       points to the input buffer
+ * @param[in]  blockSize  length of the input vector
+ * @param[out] pResult    maximum value returned here
+ * @param[out] pIndex     index of maximum value returned here
+ */
+  void arm_max_f32(
+  const float32_t * pSrc,
+        uint32_t blockSize,
+        float32_t * pResult,
+        uint32_t * pIndex);
+
+/**
+ * @brief Maximum value of absolute values of a floating-point vector.
+ * @param[in]  pSrc       points to the input buffer
+ * @param[in]  blockSize  length of the input vector
+ * @param[out] pResult    maximum value returned here
+ * @param[out] pIndex     index of maximum value returned here
+ */
+  void arm_absmax_f32(
+  const float32_t * pSrc,
+        uint32_t blockSize,
+        float32_t * pResult,
+        uint32_t * pIndex);
+
+ /**
+ * @brief Maximum value of absolute values of a floating-point vector.
+ * @param[in]  pSrc       points to the input buffer
+ * @param[in]  blockSize  length of the input vector
+ * @param[out] pResult    maximum value returned here
+ */
+  void arm_absmax_no_idx_f32(
+  const float32_t * pSrc,
+        uint32_t blockSize,
+        float32_t * pResult);
+
+/**
+ * @brief Maximum value of a floating-point vector.
+ * @param[in]  pSrc       points to the input buffer
+ * @param[in]  blockSize  length of the input vector
+ * @param[out] pResult    maximum value returned here
+ * @param[out] pIndex     index of maximum value returned here
+ */
+  void arm_max_f64(
+  const float64_t * pSrc,
+        uint32_t blockSize,
+        float64_t * pResult,
+        uint32_t * pIndex);
+
+/**
+ * @brief Maximum value of absolute values of a floating-point vector.
+ * @param[in]  pSrc       points to the input buffer
+ * @param[in]  blockSize  length of the input vector
+ * @param[out] pResult    maximum value returned here
+ * @param[out] pIndex     index of maximum value returned here
+ */
+  void arm_absmax_f64(
+  const float64_t * pSrc,
+        uint32_t blockSize,
+        float64_t * pResult,
+        uint32_t * pIndex);
+
+/**
+ * @brief Maximum value of absolute values of a floating-point vector.
+ * @param[in]  pSrc       points to the input buffer
+ * @param[in]  blockSize  length of the input vector
+ * @param[out] pResult    maximum value returned here
+ */
+  void arm_absmax_no_idx_f64(
+  const float64_t * pSrc,
+        uint32_t blockSize,
+        float64_t * pResult);
+
+  /**
+    @brief         Maximum value of a floating-point vector.
+    @param[in]     pSrc       points to the input vector
+    @param[in]     blockSize  number of samples in input vector
+    @param[out]    pResult    maximum value returned here
+    @return        none
+   */
+  void arm_max_no_idx_f32(
+      const float32_t *pSrc,
+      uint32_t   blockSize,
+      float32_t *pResult);
+
+  /**
+    @brief         Minimum value of a floating-point vector.
+    @param[in]     pSrc       points to the input vector
+    @param[in]     blockSize  number of samples in input vector
+    @param[out]    pResult    minimum value returned here
+    @return        none
+   */
+  void arm_min_no_idx_f32(
+      const float32_t *pSrc,
+      uint32_t   blockSize,
+      float32_t *pResult);
+
+  /**
+    @brief         Maximum value of a floating-point vector.
+    @param[in]     pSrc       points to the input vector
+    @param[in]     blockSize  number of samples in input vector
+    @param[out]    pResult    maximum value returned here
+    @return        none
+   */
+  void arm_max_no_idx_f64(
+      const float64_t *pSrc,
+      uint32_t   blockSize,
+      float64_t *pResult);
+
+  /**
+    @brief         Maximum value of a q31 vector.
+    @param[in]     pSrc       points to the input vector
+    @param[in]     blockSize  number of samples in input vector
+    @param[out]    pResult    maximum value returned here
+    @return        none
+   */
+  void arm_max_no_idx_q31(
+      const q31_t *pSrc,
+      uint32_t   blockSize,
+      q31_t *pResult);
+
+  /**
+    @brief         Maximum value of a q15 vector.
+    @param[in]     pSrc       points to the input vector
+    @param[in]     blockSize  number of samples in input vector
+    @param[out]    pResult    maximum value returned here
+    @return        none
+   */
+  void arm_max_no_idx_q15(
+      const q15_t *pSrc,
+      uint32_t   blockSize,
+      q15_t *pResult);
+
+  /**
+    @brief         Maximum value of a q7 vector.
+    @param[in]     pSrc       points to the input vector
+    @param[in]     blockSize  number of samples in input vector
+    @param[out]    pResult    maximum value returned here
+    @return        none
+   */
+  void arm_max_no_idx_q7(
+      const q7_t *pSrc,
+      uint32_t   blockSize,
+      q7_t *pResult);
+
+  /**
+    @brief         Minimum value of a floating-point vector.
+    @param[in]     pSrc       points to the input vector
+    @param[in]     blockSize  number of samples in input vector
+    @param[out]    pResult    minimum value returned here
+    @return        none
+   */
+  void arm_min_no_idx_f64(
+      const float64_t *pSrc,
+      uint32_t   blockSize,
+      float64_t *pResult);
+
+/**
+    @brief         Minimum value of a q31 vector.
+    @param[in]     pSrc       points to the input vector
+    @param[in]     blockSize  number of samples in input vector
+    @param[out]    pResult    minimum value returned here
+    @return        none
+   */
+  void arm_min_no_idx_q31(
+      const q31_t *pSrc,
+      uint32_t   blockSize,
+      q31_t *pResult);
+
+  /**
+    @brief         Minimum value of a q15 vector.
+    @param[in]     pSrc       points to the input vector
+    @param[in]     blockSize  number of samples in input vector
+    @param[out]    pResult    minimum value returned here
+    @return        none
+   */
+  void arm_min_no_idx_q15(
+      const q15_t *pSrc,
+      uint32_t   blockSize,
+      q15_t *pResult);
+
+  /**
+    @brief         Minimum value of a q7 vector.
+    @param[in]     pSrc       points to the input vector
+    @param[in]     blockSize  number of samples in input vector
+    @param[out]    pResult    minimum value returned here
+    @return        none
+   */
+  void arm_min_no_idx_q7(
+      const q7_t *pSrc,
+      uint32_t   blockSize,
+      q7_t *pResult);
+
+/**
+  @brief         Mean square error between two Q7 vectors.
+  @param[in]     pSrcA       points to the first input vector
+  @param[in]     pSrcB       points to the second input vector
+  @param[in]     blockSize  number of samples in input vector
+  @param[out]    pResult    mean square error
+  @return        none 
+*/
+  
+void arm_mse_q7(
+  const q7_t * pSrcA,
+  const q7_t * pSrcB,
+        uint32_t blockSize,
+        q7_t * pResult);
+
+/**
+  @brief         Mean square error between two Q15 vectors.
+  @param[in]     pSrcA       points to the first input vector
+  @param[in]     pSrcB       points to the second input vector
+  @param[in]     blockSize  number of samples in input vector
+  @param[out]    pResult    mean square error
+  @return        none 
+*/
+  
+void arm_mse_q15(
+  const q15_t * pSrcA,
+  const q15_t * pSrcB,
+        uint32_t blockSize,
+        q15_t * pResult);
+
+/**
+  @brief         Mean square error between two Q31 vectors.
+  @param[in]     pSrcA       points to the first input vector
+  @param[in]     pSrcB       points to the second input vector
+  @param[in]     blockSize  number of samples in input vector
+  @param[out]    pResult    mean square error
+  @return        none 
+*/
+  
+void arm_mse_q31(
+  const q31_t * pSrcA,
+  const q31_t * pSrcB,
+        uint32_t blockSize,
+        q31_t * pResult);
+
+/**
+  @brief         Mean square error between two single precision float vectors.
+  @param[in]     pSrcA       points to the first input vector
+  @param[in]     pSrcB       points to the second input vector
+  @param[in]     blockSize  number of samples in input vector
+  @param[out]    pResult    mean square error
+  @return        none 
+*/
+  
+void arm_mse_f32(
+  const float32_t * pSrcA,
+  const float32_t * pSrcB,
+        uint32_t blockSize,
+        float32_t * pResult);
+
+/**
+  @brief         Mean square error between two double precision float vectors.
+  @param[in]     pSrcA       points to the first input vector
+  @param[in]     pSrcB       points to the second input vector
+  @param[in]     blockSize  number of samples in input vector
+  @param[out]    pResult    mean square error
+  @return        none 
+*/
+  
+void arm_mse_f64(
+  const float64_t * pSrcA,
+  const float64_t * pSrcB,
+        uint32_t blockSize,
+        float64_t * pResult);
+
+#ifdef   __cplusplus
+}
+#endif
+
+#endif /* ifndef _STATISTICS_FUNCTIONS_H_ */

bsp/cmsis/include/dsp/statistics_functions_f16.h

@@ -0,0 +1,266 @@
+/******************************************************************************
+ * @file     statistics_functions_f16.h
+ * @brief    Public header file for CMSIS DSP Library
+ * @version  V1.10.0
+ * @date     08 July 2021
+ * Target Processor: Cortex-M and Cortex-A cores
+ ******************************************************************************/
+/*
+ * Copyright (c) 2010-2020 Arm Limited or its affiliates. All rights reserved.
+ *
+ * SPDX-License-Identifier: Apache-2.0
+ *
+ * Licensed under the Apache License, Version 2.0 (the License); you may
+ * not use this file except in compliance with the License.
+ * You may obtain a copy of the License at
+ *
+ * www.apache.org/licenses/LICENSE-2.0
+ *
+ * Unless required by applicable law or agreed to in writing, software
+ * distributed under the License is distributed on an AS IS BASIS, WITHOUT
+ * WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ * See the License for the specific language governing permissions and
+ * limitations under the License.
+ */
+
+ 
+#ifndef _STATISTICS_FUNCTIONS_F16_H_
+#define _STATISTICS_FUNCTIONS_F16_H_
+
+#include "arm_math_types_f16.h"
+#include "arm_math_memory.h"
+
+#include "dsp/none.h"
+#include "dsp/utils.h"
+
+#include "dsp/basic_math_functions_f16.h"
+#include "dsp/fast_math_functions_f16.h"
+
+#ifdef   __cplusplus
+extern "C"
+{
+#endif
+
+#if defined(ARM_FLOAT16_SUPPORTED)
+
+ /**
+   * @brief  Sum of the squares of the elements of a floating-point vector.
+   * @param[in]  pSrc       is input pointer
+   * @param[in]  blockSize  is the number of samples to process
+   * @param[out] pResult    is output value.
+   */
+  void arm_power_f16(
+  const float16_t * pSrc,
+        uint32_t blockSize,
+        float16_t * pResult);
+
+ /**
+   * @brief  Mean value of a floating-point vector.
+   * @param[in]  pSrc       is input pointer
+   * @param[in]  blockSize  is the number of samples to process
+   * @param[out] pResult    is output value.
+   */
+  void arm_mean_f16(
+  const float16_t * pSrc,
+        uint32_t blockSize,
+        float16_t * pResult);
+
+  /**
+   * @brief  Variance of the elements of a floating-point vector.
+   * @param[in]  pSrc       is input pointer
+   * @param[in]  blockSize  is the number of samples to process
+   * @param[out] pResult    is output value.
+   */
+  void arm_var_f16(
+  const float16_t * pSrc,
+        uint32_t blockSize,
+        float16_t * pResult);
+
+ /**
+   * @brief  Root Mean Square of the elements of a floating-point vector.
+   * @param[in]  pSrc       is input pointer
+   * @param[in]  blockSize  is the number of samples to process
+   * @param[out] pResult    is output value.
+   */
+  void arm_rms_f16(
+  const float16_t * pSrc,
+        uint32_t blockSize,
+        float16_t * pResult);
+
+ /**
+   * @brief  Standard deviation of the elements of a floating-point vector.
+   * @param[in]  pSrc       is input pointer
+   * @param[in]  blockSize  is the number of samples to process
+   * @param[out] pResult    is output value.
+   */
+  void arm_std_f16(
+  const float16_t * pSrc,
+        uint32_t blockSize,
+        float16_t * pResult);
+
+ /**
+   * @brief  Minimum value of a floating-point vector.
+   * @param[in]  pSrc       is input pointer
+   * @param[in]  blockSize  is the number of samples to process
+   * @param[out] pResult    is output pointer
+   * @param[out] pIndex     is the array index of the minimum value in the input buffer.
+   */
+  void arm_min_f16(
+  const float16_t * pSrc,
+        uint32_t blockSize,
+        float16_t * pResult,
+        uint32_t * pIndex);
+
+ /**
+   * @brief  Minimum value of absolute values of a floating-point vector.
+   * @param[in]  pSrc       is input pointer
+   * @param[in]  blockSize  is the number of samples to process
+   * @param[out] pResult    is output pointer
+   * @param[out] pIndex     is the array index of the minimum value in the input buffer.
+   */
+  void arm_absmin_f16(
+  const float16_t * pSrc,
+        uint32_t blockSize,
+        float16_t * pResult,
+        uint32_t * pIndex);
+
+/**
+ * @brief Maximum value of a floating-point vector.
+ * @param[in]  pSrc       points to the input buffer
+ * @param[in]  blockSize  length of the input vector
+ * @param[out] pResult    maximum value returned here
+ * @param[out] pIndex     index of maximum value returned here
+ */
+  void arm_max_f16(
+  const float16_t * pSrc,
+        uint32_t blockSize,
+        float16_t * pResult,
+        uint32_t * pIndex);
+
+/**
+ * @brief Maximum value of absolute values of a floating-point vector.
+ * @param[in]  pSrc       points to the input buffer
+ * @param[in]  blockSize  length of the input vector
+ * @param[out] pResult    maximum value returned here
+ * @param[out] pIndex     index of maximum value returned here
+ */
+  void arm_absmax_f16(
+  const float16_t * pSrc,
+        uint32_t blockSize,
+        float16_t * pResult,
+        uint32_t * pIndex);
+
+    /**
+   * @brief  Minimum value of absolute values of a floating-point vector.
+   * @param[in]  pSrc       is input pointer
+   * @param[in]  blockSize  is the number of samples to process
+   * @param[out] pResult    is output pointer
+   */
+  void arm_absmin_no_idx_f16(
+  const float16_t * pSrc,
+        uint32_t blockSize,
+        float16_t * pResult);
+
+/**
+ * @brief Maximum value of a floating-point vector.
+ * @param[in]  pSrc       points to the input buffer
+ * @param[in]  blockSize  length of the input vector
+ * @param[out] pResult    maximum value returned here
+ */
+  void arm_absmax_no_idx_f16(
+  const float16_t * pSrc,
+        uint32_t blockSize,
+        float16_t * pResult);
+
+
+/**
+ * @brief Entropy
+ *
+ * @param[in]  pSrcA        Array of input values.
+ * @param[in]  blockSize    Number of samples in the input array.
+ * @return     Entropy      -Sum(p ln p)
+ *
+ */
+
+
+float16_t arm_entropy_f16(const float16_t * pSrcA,uint32_t blockSize);
+
+float16_t arm_logsumexp_f16(const float16_t *in, uint32_t blockSize);
+
+/**
+ * @brief Dot product with log arithmetic
+ *
+ * Vectors are containing the log of the samples
+ *
+ * @param[in]       pSrcA points to the first input vector
+ * @param[in]       pSrcB points to the second input vector
+ * @param[in]       blockSize number of samples in each vector
+ * @param[in]       pTmpBuffer temporary buffer of length blockSize
+ * @return The log of the dot product .
+ *
+ */
+
+
+float16_t arm_logsumexp_dot_prod_f16(const float16_t * pSrcA,
+  const float16_t * pSrcB,
+  uint32_t blockSize,
+  float16_t *pTmpBuffer);
+
+/**
+ * @brief Kullback-Leibler
+ *
+ * @param[in]  pSrcA         Pointer to an array of input values for probability distribution A.
+ * @param[in]  pSrcB         Pointer to an array of input values for probability distribution B.
+ * @param[in]  blockSize     Number of samples in the input array.
+ * @return Kullback-Leibler  Divergence D(A || B)
+ *
+ */
+float16_t arm_kullback_leibler_f16(const float16_t * pSrcA
+  ,const float16_t * pSrcB
+  ,uint32_t blockSize);
+
+/**
+    @brief         Maximum value of a floating-point vector.
+    @param[in]     pSrc       points to the input vector
+    @param[in]     blockSize  number of samples in input vector
+    @param[out]    pResult    maximum value returned here
+    @return        none
+   */
+  void arm_max_no_idx_f16(
+      const float16_t *pSrc,
+      uint32_t   blockSize,
+      float16_t *pResult);
+
+/**
+    @brief         Minimum value of a floating-point vector.
+    @param[in]     pSrc       points to the input vector
+    @param[in]     blockSize  number of samples in input vector
+    @param[out]    pResult    minimum value returned here
+    @return        none
+   */
+  void arm_min_no_idx_f16(
+      const float16_t *pSrc,
+      uint32_t   blockSize,
+      float16_t *pResult);
+
+/**
+  @brief         Mean square error between two half precision float vectors.
+  @param[in]     pSrcA       points to the first input vector
+  @param[in]     pSrcB       points to the second input vector
+  @param[in]     blockSize  number of samples in input vector
+  @param[out]    pResult    mean square error
+  @return        none 
+*/
+  
+void arm_mse_f16(
+  const float16_t * pSrcA,
+  const float16_t * pSrcB,
+        uint32_t blockSize,
+        float16_t * pResult);
+
+#endif /*defined(ARM_FLOAT16_SUPPORTED)*/
+#ifdef   __cplusplus
+}
+#endif
+
+#endif /* ifndef _STATISTICS_FUNCTIONS_F16_H_ */

bsp/cmsis/include/dsp/support_functions.h

@@ -0,0 +1,453 @@
+/******************************************************************************
+ * @file     support_functions.h
+ * @brief    Public header file for CMSIS DSP Library
+ * @version  V1.10.0
+ * @date     08 July 2021
+ * Target Processor: Cortex-M and Cortex-A cores
+ ******************************************************************************/
+/*
+ * Copyright (c) 2010-2020 Arm Limited or its affiliates. All rights reserved.
+ *
+ * SPDX-License-Identifier: Apache-2.0
+ *
+ * Licensed under the Apache License, Version 2.0 (the License); you may
+ * not use this file except in compliance with the License.
+ * You may obtain a copy of the License at
+ *
+ * www.apache.org/licenses/LICENSE-2.0
+ *
+ * Unless required by applicable law or agreed to in writing, software
+ * distributed under the License is distributed on an AS IS BASIS, WITHOUT
+ * WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ * See the License for the specific language governing permissions and
+ * limitations under the License.
+ */
+
+ 
+#ifndef _SUPPORT_FUNCTIONS_H_
+#define _SUPPORT_FUNCTIONS_H_
+
+#include "arm_math_types.h"
+#include "arm_math_memory.h"
+
+#include "dsp/none.h"
+#include "dsp/utils.h"
+
+#ifdef   __cplusplus
+extern "C"
+{
+#endif
+
+/**
+ * @defgroup groupSupport Support Functions
+ */
+
+
+/**
+   * @brief Converts the elements of the floating-point vector to Q31 vector.
+   * @param[in]  pSrc       points to the floating-point input vector
+   * @param[out] pDst       points to the Q31 output vector
+   * @param[in]  blockSize  length of the input vector
+   */
+  void arm_float_to_q31(
+  const float32_t * pSrc,
+        q31_t * pDst,
+        uint32_t blockSize);
+
+
+  /**
+   * @brief Converts the elements of the floating-point vector to Q15 vector.
+   * @param[in]  pSrc       points to the floating-point input vector
+   * @param[out] pDst       points to the Q15 output vector
+   * @param[in]  blockSize  length of the input vector
+   */
+  void arm_float_to_q15(
+  const float32_t * pSrc,
+        q15_t * pDst,
+        uint32_t blockSize);
+
+
+  /**
+   * @brief Converts the elements of the floating-point vector to Q7 vector.
+   * @param[in]  pSrc       points to the floating-point input vector
+   * @param[out] pDst       points to the Q7 output vector
+   * @param[in]  blockSize  length of the input vector
+   */
+  void arm_float_to_q7(
+  const float32_t * pSrc,
+        q7_t * pDst,
+        uint32_t blockSize);
+
+
+  /**
+   * @brief  Converts the elements of the Q31 vector to floating-point vector.
+   * @param[in]  pSrc       is input pointer
+   * @param[out] pDst       is output pointer
+   * @param[in]  blockSize  is the number of samples to process
+   */
+  void arm_q31_to_float(
+  const q31_t * pSrc,
+        float32_t * pDst,
+        uint32_t blockSize);
+
+
+  /**
+   * @brief  Converts the elements of the Q31 vector to Q15 vector.
+   * @param[in]  pSrc       is input pointer
+   * @param[out] pDst       is output pointer
+   * @param[in]  blockSize  is the number of samples to process
+   */
+  void arm_q31_to_q15(
+  const q31_t * pSrc,
+        q15_t * pDst,
+        uint32_t blockSize);
+
+
+  /**
+   * @brief  Converts the elements of the Q31 vector to Q7 vector.
+   * @param[in]  pSrc       is input pointer
+   * @param[out] pDst       is output pointer
+   * @param[in]  blockSize  is the number of samples to process
+   */
+  void arm_q31_to_q7(
+  const q31_t * pSrc,
+        q7_t * pDst,
+        uint32_t blockSize);
+
+
+  /**
+   * @brief  Converts the elements of the Q15 vector to floating-point vector.
+   * @param[in]  pSrc       is input pointer
+   * @param[out] pDst       is output pointer
+   * @param[in]  blockSize  is the number of samples to process
+   */
+  void arm_q15_to_float(
+  const q15_t * pSrc,
+        float32_t * pDst,
+        uint32_t blockSize);
+
+
+  /**
+   * @brief  Converts the elements of the Q15 vector to Q31 vector.
+   * @param[in]  pSrc       is input pointer
+   * @param[out] pDst       is output pointer
+   * @param[in]  blockSize  is the number of samples to process
+   */
+  void arm_q15_to_q31(
+  const q15_t * pSrc,
+        q31_t * pDst,
+        uint32_t blockSize);
+
+
+  /**
+   * @brief  Converts the elements of the Q15 vector to Q7 vector.
+   * @param[in]  pSrc       is input pointer
+   * @param[out] pDst       is output pointer
+   * @param[in]  blockSize  is the number of samples to process
+   */
+  void arm_q15_to_q7(
+  const q15_t * pSrc,
+        q7_t * pDst,
+        uint32_t blockSize);
+
+
+  /**
+   * @brief  Converts the elements of the Q7 vector to floating-point vector.
+   * @param[in]  pSrc       is input pointer
+   * @param[out] pDst       is output pointer
+   * @param[in]  blockSize  is the number of samples to process
+   */
+  void arm_q7_to_float(
+  const q7_t * pSrc,
+        float32_t * pDst,
+        uint32_t blockSize);
+
+
+  /**
+   * @brief  Converts the elements of the Q7 vector to Q31 vector.
+   * @param[in]  pSrc       input pointer
+   * @param[out] pDst       output pointer
+   * @param[in]  blockSize  number of samples to process
+   */
+  void arm_q7_to_q31(
+  const q7_t * pSrc,
+        q31_t * pDst,
+        uint32_t blockSize);
+
+
+  /**
+   * @brief  Converts the elements of the Q7 vector to Q15 vector.
+   * @param[in]  pSrc       input pointer
+   * @param[out] pDst       output pointer
+   * @param[in]  blockSize  number of samples to process
+   */
+  void arm_q7_to_q15(
+  const q7_t * pSrc,
+        q15_t * pDst,
+        uint32_t blockSize);
+
+
+
+
+  
+  /**
+   * @brief Struct for specifying sorting algorithm
+   */
+  typedef enum
+  {
+    ARM_SORT_BITONIC   = 0,
+             /**< Bitonic sort   */
+    ARM_SORT_BUBBLE    = 1,
+             /**< Bubble sort    */
+    ARM_SORT_HEAP      = 2,
+             /**< Heap sort      */
+    ARM_SORT_INSERTION = 3,
+             /**< Insertion sort */
+    ARM_SORT_QUICK     = 4,
+             /**< Quick sort     */
+    ARM_SORT_SELECTION = 5
+             /**< Selection sort */
+  } arm_sort_alg;
+
+  /**
+   * @brief Struct for specifying sorting algorithm
+   */
+  typedef enum
+  {
+    ARM_SORT_DESCENDING = 0,
+             /**< Descending order (9 to 0) */
+    ARM_SORT_ASCENDING = 1
+             /**< Ascending order (0 to 9) */
+  } arm_sort_dir;
+
+  /**
+   * @brief Instance structure for the sorting algorithms.
+   */
+  typedef struct            
+  {
+    arm_sort_alg alg;        /**< Sorting algorithm selected */
+    arm_sort_dir dir;        /**< Sorting order (direction)  */
+  } arm_sort_instance_f32;  
+
+  /**
+   * @param[in]  S          points to an instance of the sorting structure.
+   * @param[in]  pSrc       points to the block of input data.
+   * @param[out] pDst       points to the block of output data.
+   * @param[in]  blockSize  number of samples to process.
+   */
+  void arm_sort_f32(
+    const arm_sort_instance_f32 * S, 
+          float32_t * pSrc, 
+          float32_t * pDst, 
+          uint32_t blockSize);
+
+  /**
+   * @param[in,out]  S            points to an instance of the sorting structure.
+   * @param[in]      alg          Selected algorithm.
+   * @param[in]      dir          Sorting order.
+   */
+  void arm_sort_init_f32(
+    arm_sort_instance_f32 * S, 
+    arm_sort_alg alg, 
+    arm_sort_dir dir); 
+
+  /**
+   * @brief Instance structure for the sorting algorithms.
+   */
+  typedef struct            
+  {
+    arm_sort_dir dir;        /**< Sorting order (direction)  */
+    float32_t * buffer;      /**< Working buffer */
+  } arm_merge_sort_instance_f32;  
+
+  /**
+   * @param[in]      S          points to an instance of the sorting structure.
+   * @param[in,out]  pSrc       points to the block of input data.
+   * @param[out]     pDst       points to the block of output data
+   * @param[in]      blockSize  number of samples to process.
+   */
+  void arm_merge_sort_f32(
+    const arm_merge_sort_instance_f32 * S,
+          float32_t *pSrc,
+          float32_t *pDst,
+          uint32_t blockSize);
+
+  /**
+   * @param[in,out]  S            points to an instance of the sorting structure.
+   * @param[in]      dir          Sorting order.
+   * @param[in]      buffer       Working buffer.
+   */
+  void arm_merge_sort_init_f32(
+    arm_merge_sort_instance_f32 * S,
+    arm_sort_dir dir,
+    float32_t * buffer);
+
+ 
+ 
+  /**
+   * @brief  Copies the elements of a floating-point vector.
+   * @param[in]  pSrc       input pointer
+   * @param[out] pDst       output pointer
+   * @param[in]  blockSize  number of samples to process
+   */
+  void arm_copy_f32(
+  const float32_t * pSrc,
+        float32_t * pDst,
+        uint32_t blockSize);
+
+ 
+ 
+  /**
+   * @brief  Copies the elements of a floating-point vector.
+   * @param[in]  pSrc       input pointer
+   * @param[out] pDst       output pointer
+   * @param[in]  blockSize  number of samples to process
+   */
+  void arm_copy_f64(
+  const float64_t * pSrc,
+        float64_t * pDst,
+        uint32_t blockSize);
+
+
+
+  /**
+   * @brief  Copies the elements of a Q7 vector.
+   * @param[in]  pSrc       input pointer
+   * @param[out] pDst       output pointer
+   * @param[in]  blockSize  number of samples to process
+   */
+  void arm_copy_q7(
+  const q7_t * pSrc,
+        q7_t * pDst,
+        uint32_t blockSize);
+
+
+  /**
+   * @brief  Copies the elements of a Q15 vector.
+   * @param[in]  pSrc       input pointer
+   * @param[out] pDst       output pointer
+   * @param[in]  blockSize  number of samples to process
+   */
+  void arm_copy_q15(
+  const q15_t * pSrc,
+        q15_t * pDst,
+        uint32_t blockSize);
+
+
+  /**
+   * @brief  Copies the elements of a Q31 vector.
+   * @param[in]  pSrc       input pointer
+   * @param[out] pDst       output pointer
+   * @param[in]  blockSize  number of samples to process
+   */
+  void arm_copy_q31(
+  const q31_t * pSrc,
+        q31_t * pDst,
+        uint32_t blockSize);
+
+
+  /**
+   * @brief  Fills a constant value into a floating-point vector.
+   * @param[in]  value      input value to be filled
+   * @param[out] pDst       output pointer
+   * @param[in]  blockSize  number of samples to process
+   */
+  void arm_fill_f32(
+        float32_t value,
+        float32_t * pDst,
+        uint32_t blockSize);
+
+
+  /**
+   * @brief  Fills a constant value into a floating-point vector.
+   * @param[in]  value      input value to be filled
+   * @param[out] pDst       output pointer
+   * @param[in]  blockSize  number of samples to process
+   */
+  void arm_fill_f64(
+        float64_t value,
+        float64_t * pDst,
+        uint32_t blockSize);
+
+
+  /**
+   * @brief  Fills a constant value into a Q7 vector.
+   * @param[in]  value      input value to be filled
+   * @param[out] pDst       output pointer
+   * @param[in]  blockSize  number of samples to process
+   */
+  void arm_fill_q7(
+        q7_t value,
+        q7_t * pDst,
+        uint32_t blockSize);
+
+
+  /**
+   * @brief  Fills a constant value into a Q15 vector.
+   * @param[in]  value      input value to be filled
+   * @param[out] pDst       output pointer
+   * @param[in]  blockSize  number of samples to process
+   */
+  void arm_fill_q15(
+        q15_t value,
+        q15_t * pDst,
+        uint32_t blockSize);
+
+
+  /**
+   * @brief  Fills a constant value into a Q31 vector.
+   * @param[in]  value      input value to be filled
+   * @param[out] pDst       output pointer
+   * @param[in]  blockSize  number of samples to process
+   */
+  void arm_fill_q31(
+        q31_t value,
+        q31_t * pDst,
+        uint32_t blockSize);
+
+
+
+
+
+
+
+/**
+ * @brief Weighted sum
+ *
+ *
+ * @param[in]    *in           Array of input values.
+ * @param[in]    *weigths      Weights
+ * @param[in]    blockSize     Number of samples in the input array.
+ * @return Weighted sum
+ *
+ */
+float32_t arm_weighted_sum_f32(const float32_t *in
+  , const float32_t *weigths
+  , uint32_t blockSize);
+
+
+/**
+ * @brief Barycenter
+ *
+ *
+ * @param[in]    in         List of vectors
+ * @param[in]    weights    Weights of the vectors
+ * @param[out]   out        Barycenter
+ * @param[in]    nbVectors  Number of vectors
+ * @param[in]    vecDim     Dimension of space (vector dimension)
+ * @return       None
+ *
+ */
+void arm_barycenter_f32(const float32_t *in
+  , const float32_t *weights
+  , float32_t *out
+  , uint32_t nbVectors
+  , uint32_t vecDim);
+
+
+
+#ifdef   __cplusplus
+}
+#endif
+
+#endif /* ifndef _SUPPORT_FUNCTIONS_H_ */

bsp/cmsis/include/dsp/support_functions_f16.h

@@ -0,0 +1,187 @@
+/******************************************************************************
+ * @file     support_functions_f16.h
+ * @brief    Public header file for CMSIS DSP Library
+ * @version  V1.10.0
+ * @date     08 July 2021
+ * Target Processor: Cortex-M and Cortex-A cores
+ ******************************************************************************/
+/*
+ * Copyright (c) 2010-2020 Arm Limited or its affiliates. All rights reserved.
+ *
+ * SPDX-License-Identifier: Apache-2.0
+ *
+ * Licensed under the Apache License, Version 2.0 (the License); you may
+ * not use this file except in compliance with the License.
+ * You may obtain a copy of the License at
+ *
+ * www.apache.org/licenses/LICENSE-2.0
+ *
+ * Unless required by applicable law or agreed to in writing, software
+ * distributed under the License is distributed on an AS IS BASIS, WITHOUT
+ * WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ * See the License for the specific language governing permissions and
+ * limitations under the License.
+ */
+
+ 
+#ifndef _SUPPORT_FUNCTIONS_F16_H_
+#define _SUPPORT_FUNCTIONS_F16_H_
+
+#include "arm_math_types_f16.h"
+#include "arm_math_memory.h"
+
+#include "dsp/none.h"
+#include "dsp/utils.h"
+
+#ifdef   __cplusplus
+extern "C"
+{
+#endif
+
+#if defined(ARM_FLOAT16_SUPPORTED)
+
+  /**
+   * @brief  Copies the elements of a floating-point vector.
+   * @param[in]  pSrc       input pointer
+   * @param[out] pDst       output pointer
+   * @param[in]  blockSize  number of samples to process
+   */
+void arm_copy_f16(const float16_t * pSrc, float16_t * pDst, uint32_t blockSize);
+
+  /**
+   * @brief  Fills a constant value into a floating-point vector.
+   * @param[in]  value      input value to be filled
+   * @param[out] pDst       output pointer
+   * @param[in]  blockSize  number of samples to process
+   */
+void arm_fill_f16(float16_t value, float16_t * pDst, uint32_t blockSize);
+
+/**
+   * @brief Converts the elements of the floating-point vector to Q31 vector.
+   * @param[in]  pSrc       points to the f16 input vector
+   * @param[out] pDst       points to the q15 output vector
+   * @param[in]  blockSize  length of the input vector
+   */
+void arm_f16_to_q15(const float16_t * pSrc, q15_t * pDst, uint32_t blockSize);
+
+/**
+   * @brief Converts the elements of the floating-point vector to Q31 vector.
+   * @param[in]  pSrc       points to the q15 input vector
+   * @param[out] pDst       points to the f16 output vector
+   * @param[in]  blockSize  length of the input vector
+   */
+void arm_q15_to_f16(const q15_t * pSrc, float16_t * pDst, uint32_t blockSize);
+
+
+/**
+   * @brief Converts the elements of the floating-point vector to Q31 vector.
+   * @param[in]  pSrc       points to the f32 input vector
+   * @param[out] pDst       points to the f16 output vector
+   * @param[in]  blockSize  length of the input vector
+   */
+void arm_float_to_f16(const float32_t * pSrc, float16_t * pDst, uint32_t blockSize);
+
+/**
+   * @brief Converts the elements of the floating-point vector to Q31 vector.
+   * @param[in]  pSrc       points to the f16 input vector
+   * @param[out] pDst       points to the f32 output vector
+   * @param[in]  blockSize  length of the input vector
+   */
+void arm_f16_to_float(const float16_t * pSrc, float32_t * pDst, uint32_t blockSize);
+
+/**
+ * @brief Weighted sum
+ *
+ *
+ * @param[in]    *in           Array of input values.
+ * @param[in]    *weigths      Weights
+ * @param[in]    blockSize     Number of samples in the input array.
+ * @return Weighted sum
+ *
+ */
+float16_t arm_weighted_sum_f16(const float16_t *in
+  , const float16_t *weigths
+  , uint32_t blockSize);
+
+/**
+ * @brief Barycenter
+ *
+ *
+ * @param[in]    in         List of vectors
+ * @param[in]    weights    Weights of the vectors
+ * @param[out]   out        Barycenter
+ * @param[in]    nbVectors  Number of vectors
+ * @param[in]    vecDim     Dimension of space (vector dimension)
+ * @return       None
+ *
+ */
+void arm_barycenter_f16(const float16_t *in
+  , const float16_t *weights
+  , float16_t *out
+  , uint32_t nbVectors
+  , uint32_t vecDim);
+
+
+/**
+  @ingroup groupSupport
+ */
+
+/**
+ * @defgroup typecast Typecasting
+ */
+
+/**
+  @addtogroup typecast
+  @{
+ */
+
+/**
+   * @brief  Interpret a f16 as an s16 value
+   * @param[in] x  input value.
+   * @return  return value.
+   * 
+   * @par    Description
+   *            It is a typecast. No conversion of the float to int is done.
+   *            The memcpy will be optimized out by the compiler.
+   *            memcpy is used to prevent type punning issues.
+   *            With gcc, -fno-builtins MUST not be used or the
+   *            memcpy will not be optimized out.
+   */
+__STATIC_INLINE int16_t arm_typecast_s16_f16(float16_t x)
+{
+   int16_t res;
+   res=*(int16_t*)memcpy((char*)&res,(char*)&x,sizeof(float16_t));
+   return(res);
+}
+
+/**
+   * @brief  Interpret an s16 as an f16 value
+   * @param[in] x  input value.
+   * @return  return value.
+   * 
+   * @par    Description
+   *            It is a typecast. No conversion of the int to float is done.
+   *            The memcpy will be optimized out by the compiler.
+   *            memcpy is used to prevent type punning issues.
+   *            With gcc, -fno-builtins MUST not be used or the
+   *            memcpy will not be optimized out.
+   */
+__STATIC_INLINE float16_t arm_typecast_f16_s16(int16_t x)
+{
+   float16_t res;
+   res=*(float16_t*)memcpy((char*)&res,(char*)&x,sizeof(int16_t));
+   return(res);
+}
+
+
+/**
+  @} end of typecast group
+ */
+
+
+#endif /*defined(ARM_FLOAT16_SUPPORTED)*/
+#ifdef   __cplusplus
+}
+#endif
+
+#endif /* ifndef _SUPPORT_FUNCTIONS_F16_H_ */

bsp/cmsis/include/dsp/svm_defines.h

@@ -0,0 +1,46 @@
+/******************************************************************************
+ * @file     svm_defines.h
+ * @brief    Public header file for CMSIS DSP Library
+ * @version  V1.10.0
+ * @date     08 July 2021
+ *
+ * Target Processor: Cortex-M and Cortex-A cores
+ ******************************************************************************/
+/*
+ * Copyright (c) 2010-2020 Arm Limited or its affiliates. All rights reserved.
+ *
+ * SPDX-License-Identifier: Apache-2.0
+ *
+ * Licensed under the Apache License, Version 2.0 (the License); you may
+ * not use this file except in compliance with the License.
+ * You may obtain a copy of the License at
+ *
+ * www.apache.org/licenses/LICENSE-2.0
+ *
+ * Unless required by applicable law or agreed to in writing, software
+ * distributed under the License is distributed on an AS IS BASIS, WITHOUT
+ * WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ * See the License for the specific language governing permissions and
+ * limitations under the License.
+ */
+
+ 
+#ifndef _SVM_DEFINES_H_
+#define _SVM_DEFINES_H_
+
+/**
+ * @brief Struct for specifying SVM Kernel
+ */
+typedef enum
+{
+    ARM_ML_KERNEL_LINEAR = 0,
+             /**< Linear kernel */
+    ARM_ML_KERNEL_POLYNOMIAL = 1,
+             /**< Polynomial kernel */
+    ARM_ML_KERNEL_RBF = 2,
+             /**< Radial Basis Function kernel */
+    ARM_ML_KERNEL_SIGMOID = 3
+             /**< Sigmoid kernel */
+} arm_ml_kernel_type;
+
+#endif

bsp/cmsis/include/dsp/svm_functions.h

@@ -0,0 +1,299 @@
+/******************************************************************************
+ * @file     svm_functions.h
+ * @brief    Public header file for CMSIS DSP Library
+ * @version  V1.10.0
+ * @date     08 July 2021
+ * Target Processor: Cortex-M and Cortex-A cores
+ ******************************************************************************/
+/*
+ * Copyright (c) 2010-2020 Arm Limited or its affiliates. All rights reserved.
+ *
+ * SPDX-License-Identifier: Apache-2.0
+ *
+ * Licensed under the Apache License, Version 2.0 (the License); you may
+ * not use this file except in compliance with the License.
+ * You may obtain a copy of the License at
+ *
+ * www.apache.org/licenses/LICENSE-2.0
+ *
+ * Unless required by applicable law or agreed to in writing, software
+ * distributed under the License is distributed on an AS IS BASIS, WITHOUT
+ * WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ * See the License for the specific language governing permissions and
+ * limitations under the License.
+ */
+
+ 
+#ifndef _SVM_FUNCTIONS_H_
+#define _SVM_FUNCTIONS_H_
+
+#include "arm_math_types.h"
+#include "arm_math_memory.h"
+
+#include "dsp/none.h"
+#include "dsp/utils.h"
+#include "dsp/svm_defines.h"
+
+#ifdef   __cplusplus
+extern "C"
+{
+#endif
+
+#define STEP(x) (x) <= 0 ? 0 : 1
+
+/**
+ * @defgroup groupSVM SVM Functions
+ * This set of functions is implementing SVM classification on 2 classes.
+ * The training must be done from scikit-learn. The parameters can be easily
+ * generated from the scikit-learn object. Some examples are given in
+ * DSP/Testing/PatternGeneration/SVM.py
+ *
+ * If more than 2 classes are needed, the functions in this folder 
+ * will have to be used, as building blocks, to do multi-class classification.
+ *
+ * No multi-class classification is provided in this SVM folder.
+ * 
+ */
+
+/**
+ * @brief Integer exponentiation
+ * @param[in]    x           value
+ * @param[in]    nb          integer exponent >= 1
+ * @return x^nb
+ *
+ */
+__STATIC_INLINE float32_t arm_exponent_f32(float32_t x, int32_t nb)
+{
+    float32_t r = x;
+    nb --;
+    while(nb > 0)
+    {
+        r = r * x;
+        nb--;
+    }
+    return(r);
+}
+
+  
+
+
+
+/**
+ * @brief Instance structure for linear SVM prediction function.
+ */
+typedef struct
+{
+  uint32_t        nbOfSupportVectors;     /**< Number of support vectors */
+  uint32_t        vectorDimension;        /**< Dimension of vector space */
+  float32_t       intercept;              /**< Intercept */
+  const float32_t *dualCoefficients;      /**< Dual coefficients */
+  const float32_t *supportVectors;        /**< Support vectors */
+  const int32_t   *classes;               /**< The two SVM classes */
+} arm_svm_linear_instance_f32;
+
+
+/**
+ * @brief Instance structure for polynomial SVM prediction function.
+ */
+typedef struct
+{
+  uint32_t        nbOfSupportVectors;     /**< Number of support vectors */
+  uint32_t        vectorDimension;        /**< Dimension of vector space */
+  float32_t       intercept;              /**< Intercept */
+  const float32_t *dualCoefficients;      /**< Dual coefficients */
+  const float32_t *supportVectors;        /**< Support vectors */
+  const int32_t   *classes;               /**< The two SVM classes */
+  int32_t         degree;                 /**< Polynomial degree */
+  float32_t       coef0;                  /**< Polynomial constant */
+  float32_t       gamma;                  /**< Gamma factor */
+} arm_svm_polynomial_instance_f32;
+
+/**
+ * @brief Instance structure for rbf SVM prediction function.
+ */
+typedef struct
+{
+  uint32_t        nbOfSupportVectors;     /**< Number of support vectors */
+  uint32_t        vectorDimension;        /**< Dimension of vector space */
+  float32_t       intercept;              /**< Intercept */
+  const float32_t *dualCoefficients;      /**< Dual coefficients */
+  const float32_t *supportVectors;        /**< Support vectors */
+  const int32_t   *classes;               /**< The two SVM classes */
+  float32_t       gamma;                  /**< Gamma factor */
+} arm_svm_rbf_instance_f32;
+
+/**
+ * @brief Instance structure for sigmoid SVM prediction function.
+ */
+typedef struct
+{
+  uint32_t        nbOfSupportVectors;     /**< Number of support vectors */
+  uint32_t        vectorDimension;        /**< Dimension of vector space */
+  float32_t       intercept;              /**< Intercept */
+  const float32_t *dualCoefficients;      /**< Dual coefficients */
+  const float32_t *supportVectors;        /**< Support vectors */
+  const int32_t   *classes;               /**< The two SVM classes */
+  float32_t       coef0;                  /**< Independent constant */
+  float32_t       gamma;                  /**< Gamma factor */
+} arm_svm_sigmoid_instance_f32;
+
+/**
+ * @brief        SVM linear instance init function
+ * @param[in]    S                      Parameters for SVM functions
+ * @param[in]    nbOfSupportVectors     Number of support vectors
+ * @param[in]    vectorDimension        Dimension of vector space
+ * @param[in]    intercept              Intercept
+ * @param[in]    dualCoefficients       Array of dual coefficients
+ * @param[in]    supportVectors         Array of support vectors
+ * @param[in]    classes                Array of 2 classes ID
+ * @return none.
+ *
+ */
+
+
+void arm_svm_linear_init_f32(arm_svm_linear_instance_f32 *S, 
+  uint32_t nbOfSupportVectors,
+  uint32_t vectorDimension,
+  float32_t intercept,
+  const float32_t *dualCoefficients,
+  const float32_t *supportVectors,
+  const int32_t  *classes);
+
+/**
+ * @brief SVM linear prediction
+ * @param[in]    S          Pointer to an instance of the linear SVM structure.
+ * @param[in]    in         Pointer to input vector
+ * @param[out]   pResult    Decision value
+ * @return none.
+ *
+ */
+  
+void arm_svm_linear_predict_f32(const arm_svm_linear_instance_f32 *S, 
+   const float32_t * in, 
+   int32_t * pResult);
+
+
+/**
+ * @brief        SVM polynomial instance init function
+ * @param[in]    S                      points to an instance of the polynomial SVM structure.
+ * @param[in]    nbOfSupportVectors     Number of support vectors
+ * @param[in]    vectorDimension        Dimension of vector space
+ * @param[in]    intercept              Intercept
+ * @param[in]    dualCoefficients       Array of dual coefficients
+ * @param[in]    supportVectors         Array of support vectors
+ * @param[in]    classes                Array of 2 classes ID
+ * @param[in]    degree                 Polynomial degree
+ * @param[in]    coef0                  coeff0 (scikit-learn terminology)
+ * @param[in]    gamma                  gamma (scikit-learn terminology)
+ * @return none.
+ *
+ */
+
+
+void arm_svm_polynomial_init_f32(arm_svm_polynomial_instance_f32 *S, 
+  uint32_t nbOfSupportVectors,
+  uint32_t vectorDimension,
+  float32_t intercept,
+  const float32_t *dualCoefficients,
+  const float32_t *supportVectors,
+  const int32_t   *classes,
+  int32_t      degree,
+  float32_t coef0,
+  float32_t gamma
+  );
+
+/**
+ * @brief SVM polynomial prediction
+ * @param[in]    S          Pointer to an instance of the polynomial SVM structure.
+ * @param[in]    in         Pointer to input vector
+ * @param[out]   pResult    Decision value
+ * @return none.
+ *
+ */
+void arm_svm_polynomial_predict_f32(const arm_svm_polynomial_instance_f32 *S, 
+   const float32_t * in, 
+   int32_t * pResult);
+
+
+/**
+ * @brief        SVM radial basis function instance init function
+ * @param[in]    S                      points to an instance of the polynomial SVM structure.
+ * @param[in]    nbOfSupportVectors     Number of support vectors
+ * @param[in]    vectorDimension        Dimension of vector space
+ * @param[in]    intercept              Intercept
+ * @param[in]    dualCoefficients       Array of dual coefficients
+ * @param[in]    supportVectors         Array of support vectors
+ * @param[in]    classes                Array of 2 classes ID
+ * @param[in]    gamma                  gamma (scikit-learn terminology)
+ * @return none.
+ *
+ */
+
+void arm_svm_rbf_init_f32(arm_svm_rbf_instance_f32 *S, 
+  uint32_t nbOfSupportVectors,
+  uint32_t vectorDimension,
+  float32_t intercept,
+  const float32_t *dualCoefficients,
+  const float32_t *supportVectors,
+  const int32_t   *classes,
+  float32_t gamma
+  );
+
+/**
+ * @brief SVM rbf prediction
+ * @param[in]    S         Pointer to an instance of the rbf SVM structure.
+ * @param[in]    in        Pointer to input vector
+ * @param[out]   pResult   decision value
+ * @return none.
+ *
+ */
+void arm_svm_rbf_predict_f32(const arm_svm_rbf_instance_f32 *S, 
+   const float32_t * in, 
+   int32_t * pResult);
+
+/**
+ * @brief        SVM sigmoid instance init function
+ * @param[in]    S                      points to an instance of the rbf SVM structure.
+ * @param[in]    nbOfSupportVectors     Number of support vectors
+ * @param[in]    vectorDimension        Dimension of vector space
+ * @param[in]    intercept              Intercept
+ * @param[in]    dualCoefficients       Array of dual coefficients
+ * @param[in]    supportVectors         Array of support vectors
+ * @param[in]    classes                Array of 2 classes ID
+ * @param[in]    coef0                  coeff0 (scikit-learn terminology)
+ * @param[in]    gamma                  gamma (scikit-learn terminology)
+ * @return none.
+ *
+ */
+
+void arm_svm_sigmoid_init_f32(arm_svm_sigmoid_instance_f32 *S, 
+  uint32_t nbOfSupportVectors,
+  uint32_t vectorDimension,
+  float32_t intercept,
+  const float32_t *dualCoefficients,
+  const float32_t *supportVectors,
+  const int32_t   *classes,
+  float32_t coef0,
+  float32_t gamma
+  );
+
+/**
+ * @brief SVM sigmoid prediction
+ * @param[in]    S        Pointer to an instance of the rbf SVM structure.
+ * @param[in]    in       Pointer to input vector
+ * @param[out]   pResult  Decision value
+ * @return none.
+ *
+ */
+void arm_svm_sigmoid_predict_f32(const arm_svm_sigmoid_instance_f32 *S, 
+   const float32_t * in, 
+   int32_t * pResult);
+
+
+
+
+#ifdef   __cplusplus
+}
+#endif
+
+#endif /* ifndef _SVM_FUNCTIONS_H_ */

bsp/cmsis/include/dsp/svm_functions_f16.h

@@ -0,0 +1,281 @@
+/******************************************************************************
+ * @file     svm_functions_f16.h
+ * @brief    Public header file for CMSIS DSP Library
+ * @version  V1.10.0
+ * @date     08 July 2021
+ * Target Processor: Cortex-M and Cortex-A cores
+ ******************************************************************************/
+/*
+ * Copyright (c) 2010-2020 Arm Limited or its affiliates. All rights reserved.
+ *
+ * SPDX-License-Identifier: Apache-2.0
+ *
+ * Licensed under the Apache License, Version 2.0 (the License); you may
+ * not use this file except in compliance with the License.
+ * You may obtain a copy of the License at
+ *
+ * www.apache.org/licenses/LICENSE-2.0
+ *
+ * Unless required by applicable law or agreed to in writing, software
+ * distributed under the License is distributed on an AS IS BASIS, WITHOUT
+ * WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ * See the License for the specific language governing permissions and
+ * limitations under the License.
+ */
+
+ 
+#ifndef _SVM_FUNCTIONS_F16_H_
+#define _SVM_FUNCTIONS_F16_H_
+
+#include "arm_math_types_f16.h"
+#include "arm_math_memory.h"
+
+#include "dsp/none.h"
+#include "dsp/utils.h"
+#include "dsp/svm_defines.h"
+
+
+#ifdef   __cplusplus
+extern "C"
+{
+#endif
+
+#if defined(ARM_FLOAT16_SUPPORTED)
+
+#define STEP(x) (x) <= 0 ? 0 : 1
+
+/**
+ * @defgroup groupSVM SVM Functions
+ * This set of functions is implementing SVM classification on 2 classes.
+ * The training must be done from scikit-learn. The parameters can be easily
+ * generated from the scikit-learn object. Some examples are given in
+ * DSP/Testing/PatternGeneration/SVM.py
+ *
+ * If more than 2 classes are needed, the functions in this folder 
+ * will have to be used, as building blocks, to do multi-class classification.
+ *
+ * No multi-class classification is provided in this SVM folder.
+ * 
+ */
+
+
+
+/**
+ * @brief Instance structure for linear SVM prediction function.
+ */
+typedef struct
+{
+  uint32_t        nbOfSupportVectors;     /**< Number of support vectors */
+  uint32_t        vectorDimension;        /**< Dimension of vector space */
+  float16_t       intercept;              /**< Intercept */
+  const float16_t *dualCoefficients;      /**< Dual coefficients */
+  const float16_t *supportVectors;        /**< Support vectors */
+  const int32_t   *classes;               /**< The two SVM classes */
+} arm_svm_linear_instance_f16;
+
+
+/**
+ * @brief Instance structure for polynomial SVM prediction function.
+ */
+typedef struct
+{
+  uint32_t        nbOfSupportVectors;     /**< Number of support vectors */
+  uint32_t        vectorDimension;        /**< Dimension of vector space */
+  float16_t       intercept;              /**< Intercept */
+  const float16_t *dualCoefficients;      /**< Dual coefficients */
+  const float16_t *supportVectors;        /**< Support vectors */
+  const int32_t   *classes;               /**< The two SVM classes */
+  int32_t         degree;                 /**< Polynomial degree */
+  float16_t       coef0;                  /**< Polynomial constant */
+  float16_t       gamma;                  /**< Gamma factor */
+} arm_svm_polynomial_instance_f16;
+
+/**
+ * @brief Instance structure for rbf SVM prediction function.
+ */
+typedef struct
+{
+  uint32_t        nbOfSupportVectors;     /**< Number of support vectors */
+  uint32_t        vectorDimension;        /**< Dimension of vector space */
+  float16_t       intercept;              /**< Intercept */
+  const float16_t *dualCoefficients;      /**< Dual coefficients */
+  const float16_t *supportVectors;        /**< Support vectors */
+  const int32_t   *classes;               /**< The two SVM classes */
+  float16_t       gamma;                  /**< Gamma factor */
+} arm_svm_rbf_instance_f16;
+
+/**
+ * @brief Instance structure for sigmoid SVM prediction function.
+ */
+typedef struct
+{
+  uint32_t        nbOfSupportVectors;     /**< Number of support vectors */
+  uint32_t        vectorDimension;        /**< Dimension of vector space */
+  float16_t       intercept;              /**< Intercept */
+  const float16_t *dualCoefficients;      /**< Dual coefficients */
+  const float16_t *supportVectors;        /**< Support vectors */
+  const int32_t   *classes;               /**< The two SVM classes */
+  float16_t       coef0;                  /**< Independent constant */
+  float16_t       gamma;                  /**< Gamma factor */
+} arm_svm_sigmoid_instance_f16;
+
+/**
+ * @brief        SVM linear instance init function
+ * @param[in]    S                      Parameters for SVM functions
+ * @param[in]    nbOfSupportVectors     Number of support vectors
+ * @param[in]    vectorDimension        Dimension of vector space
+ * @param[in]    intercept              Intercept
+ * @param[in]    dualCoefficients       Array of dual coefficients
+ * @param[in]    supportVectors         Array of support vectors
+ * @param[in]    classes                Array of 2 classes ID
+ * @return none.
+ *
+ */
+
+
+void arm_svm_linear_init_f16(arm_svm_linear_instance_f16 *S, 
+  uint32_t nbOfSupportVectors,
+  uint32_t vectorDimension,
+  float16_t intercept,
+  const float16_t *dualCoefficients,
+  const float16_t *supportVectors,
+  const int32_t  *classes);
+
+/**
+ * @brief SVM linear prediction
+ * @param[in]    S          Pointer to an instance of the linear SVM structure.
+ * @param[in]    in         Pointer to input vector
+ * @param[out]   pResult    Decision value
+ * @return none.
+ *
+ */
+  
+void arm_svm_linear_predict_f16(const arm_svm_linear_instance_f16 *S, 
+   const float16_t * in, 
+   int32_t * pResult);
+
+
+/**
+ * @brief        SVM polynomial instance init function
+ * @param[in]    S                      points to an instance of the polynomial SVM structure.
+ * @param[in]    nbOfSupportVectors     Number of support vectors
+ * @param[in]    vectorDimension        Dimension of vector space
+ * @param[in]    intercept              Intercept
+ * @param[in]    dualCoefficients       Array of dual coefficients
+ * @param[in]    supportVectors         Array of support vectors
+ * @param[in]    classes                Array of 2 classes ID
+ * @param[in]    degree                 Polynomial degree
+ * @param[in]    coef0                  coeff0 (scikit-learn terminology)
+ * @param[in]    gamma                  gamma (scikit-learn terminology)
+ * @return none.
+ *
+ */
+
+
+void arm_svm_polynomial_init_f16(arm_svm_polynomial_instance_f16 *S, 
+  uint32_t nbOfSupportVectors,
+  uint32_t vectorDimension,
+  float16_t intercept,
+  const float16_t *dualCoefficients,
+  const float16_t *supportVectors,
+  const int32_t   *classes,
+  int32_t      degree,
+  float16_t coef0,
+  float16_t gamma
+  );
+
+/**
+ * @brief SVM polynomial prediction
+ * @param[in]    S          Pointer to an instance of the polynomial SVM structure.
+ * @param[in]    in         Pointer to input vector
+ * @param[out]   pResult    Decision value
+ * @return none.
+ *
+ */
+void arm_svm_polynomial_predict_f16(const arm_svm_polynomial_instance_f16 *S, 
+   const float16_t * in, 
+   int32_t * pResult);
+
+
+/**
+ * @brief        SVM radial basis function instance init function
+ * @param[in]    S                      points to an instance of the polynomial SVM structure.
+ * @param[in]    nbOfSupportVectors     Number of support vectors
+ * @param[in]    vectorDimension        Dimension of vector space
+ * @param[in]    intercept              Intercept
+ * @param[in]    dualCoefficients       Array of dual coefficients
+ * @param[in]    supportVectors         Array of support vectors
+ * @param[in]    classes                Array of 2 classes ID
+ * @param[in]    gamma                  gamma (scikit-learn terminology)
+ * @return none.
+ *
+ */
+
+void arm_svm_rbf_init_f16(arm_svm_rbf_instance_f16 *S, 
+  uint32_t nbOfSupportVectors,
+  uint32_t vectorDimension,
+  float16_t intercept,
+  const float16_t *dualCoefficients,
+  const float16_t *supportVectors,
+  const int32_t   *classes,
+  float16_t gamma
+  );
+
+/**
+ * @brief SVM rbf prediction
+ * @param[in]    S         Pointer to an instance of the rbf SVM structure.
+ * @param[in]    in        Pointer to input vector
+ * @param[out]   pResult   decision value
+ * @return none.
+ *
+ */
+void arm_svm_rbf_predict_f16(const arm_svm_rbf_instance_f16 *S, 
+   const float16_t * in, 
+   int32_t * pResult);
+
+/**
+ * @brief        SVM sigmoid instance init function
+ * @param[in]    S                      points to an instance of the rbf SVM structure.
+ * @param[in]    nbOfSupportVectors     Number of support vectors
+ * @param[in]    vectorDimension        Dimension of vector space
+ * @param[in]    intercept              Intercept
+ * @param[in]    dualCoefficients       Array of dual coefficients
+ * @param[in]    supportVectors         Array of support vectors
+ * @param[in]    classes                Array of 2 classes ID
+ * @param[in]    coef0                  coeff0 (scikit-learn terminology)
+ * @param[in]    gamma                  gamma (scikit-learn terminology)
+ * @return none.
+ *
+ */
+
+void arm_svm_sigmoid_init_f16(arm_svm_sigmoid_instance_f16 *S, 
+  uint32_t nbOfSupportVectors,
+  uint32_t vectorDimension,
+  float16_t intercept,
+  const float16_t *dualCoefficients,
+  const float16_t *supportVectors,
+  const int32_t   *classes,
+  float16_t coef0,
+  float16_t gamma
+  );
+
+/**
+ * @brief SVM sigmoid prediction
+ * @param[in]    S        Pointer to an instance of the rbf SVM structure.
+ * @param[in]    in       Pointer to input vector
+ * @param[out]   pResult  Decision value
+ * @return none.
+ *
+ */
+void arm_svm_sigmoid_predict_f16(const arm_svm_sigmoid_instance_f16 *S, 
+   const float16_t * in, 
+   int32_t * pResult);
+
+
+
+#endif /*defined(ARM_FLOAT16_SUPPORTED)*/
+#ifdef   __cplusplus
+}
+#endif
+
+#endif /* ifndef _SVM_FUNCTIONS_F16_H_ */

bsp/cmsis/include/dsp/transform_functions.h

@@ -0,0 +1,735 @@
+/******************************************************************************
+ * @file     transform_functions.h
+ * @brief    Public header file for CMSIS DSP Library
+ * @version  V1.10.0
+ * @date     08 July 2021
+ * Target Processor: Cortex-M and Cortex-A cores
+ ******************************************************************************/
+/*
+ * Copyright (c) 2010-2020 Arm Limited or its affiliates. All rights reserved.
+ *
+ * SPDX-License-Identifier: Apache-2.0
+ *
+ * Licensed under the Apache License, Version 2.0 (the License); you may
+ * not use this file except in compliance with the License.
+ * You may obtain a copy of the License at
+ *
+ * www.apache.org/licenses/LICENSE-2.0
+ *
+ * Unless required by applicable law or agreed to in writing, software
+ * distributed under the License is distributed on an AS IS BASIS, WITHOUT
+ * WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ * See the License for the specific language governing permissions and
+ * limitations under the License.
+ */
+
+ 
+#ifndef _TRANSFORM_FUNCTIONS_H_
+#define _TRANSFORM_FUNCTIONS_H_
+
+#include "arm_math_types.h"
+#include "arm_math_memory.h"
+
+#include "dsp/none.h"
+#include "dsp/utils.h"
+
+#include "dsp/basic_math_functions.h"
+#include "dsp/complex_math_functions.h"
+
+#ifdef   __cplusplus
+extern "C"
+{
+#endif
+
+
+/**
+ * @defgroup groupTransforms Transform Functions
+ */
+
+
+  /**
+   * @brief Instance structure for the Q15 CFFT/CIFFT function.
+   */
+  typedef struct
+  {
+          uint16_t fftLen;                 /**< length of the FFT. */
+          uint8_t ifftFlag;                /**< flag that selects forward (ifftFlag=0) or inverse (ifftFlag=1) transform. */
+          uint8_t bitReverseFlag;          /**< flag that enables (bitReverseFlag=1) or disables (bitReverseFlag=0) bit reversal of output. */
+    const q15_t *pTwiddle;                 /**< points to the Sin twiddle factor table. */
+    const uint16_t *pBitRevTable;          /**< points to the bit reversal table. */
+          uint16_t twidCoefModifier;       /**< twiddle coefficient modifier that supports different size FFTs with the same twiddle factor table. */
+          uint16_t bitRevFactor;           /**< bit reversal modifier that supports different size FFTs with the same bit reversal table. */
+  } arm_cfft_radix2_instance_q15;
+
+/* Deprecated */
+  arm_status arm_cfft_radix2_init_q15(
+        arm_cfft_radix2_instance_q15 * S,
+        uint16_t fftLen,
+        uint8_t ifftFlag,
+        uint8_t bitReverseFlag);
+
+/* Deprecated */
+  void arm_cfft_radix2_q15(
+  const arm_cfft_radix2_instance_q15 * S,
+        q15_t * pSrc);
+
+
+  /**
+   * @brief Instance structure for the Q15 CFFT/CIFFT function.
+   */
+  typedef struct
+  {
+          uint16_t fftLen;                 /**< length of the FFT. */
+          uint8_t ifftFlag;                /**< flag that selects forward (ifftFlag=0) or inverse (ifftFlag=1) transform. */
+          uint8_t bitReverseFlag;          /**< flag that enables (bitReverseFlag=1) or disables (bitReverseFlag=0) bit reversal of output. */
+    const q15_t *pTwiddle;                 /**< points to the twiddle factor table. */
+    const uint16_t *pBitRevTable;          /**< points to the bit reversal table. */
+          uint16_t twidCoefModifier;       /**< twiddle coefficient modifier that supports different size FFTs with the same twiddle factor table. */
+          uint16_t bitRevFactor;           /**< bit reversal modifier that supports different size FFTs with the same bit reversal table. */
+  } arm_cfft_radix4_instance_q15;
+
+/* Deprecated */
+  arm_status arm_cfft_radix4_init_q15(
+        arm_cfft_radix4_instance_q15 * S,
+        uint16_t fftLen,
+        uint8_t ifftFlag,
+        uint8_t bitReverseFlag);
+
+/* Deprecated */
+  void arm_cfft_radix4_q15(
+  const arm_cfft_radix4_instance_q15 * S,
+        q15_t * pSrc);
+
+  /**
+   * @brief Instance structure for the Radix-2 Q31 CFFT/CIFFT function.
+   */
+  typedef struct
+  {
+          uint16_t fftLen;                 /**< length of the FFT. */
+          uint8_t ifftFlag;                /**< flag that selects forward (ifftFlag=0) or inverse (ifftFlag=1) transform. */
+          uint8_t bitReverseFlag;          /**< flag that enables (bitReverseFlag=1) or disables (bitReverseFlag=0) bit reversal of output. */
+    const q31_t *pTwiddle;                 /**< points to the Twiddle factor table. */
+    const uint16_t *pBitRevTable;          /**< points to the bit reversal table. */
+          uint16_t twidCoefModifier;       /**< twiddle coefficient modifier that supports different size FFTs with the same twiddle factor table. */
+          uint16_t bitRevFactor;           /**< bit reversal modifier that supports different size FFTs with the same bit reversal table. */
+  } arm_cfft_radix2_instance_q31;
+
+/* Deprecated */
+  arm_status arm_cfft_radix2_init_q31(
+        arm_cfft_radix2_instance_q31 * S,
+        uint16_t fftLen,
+        uint8_t ifftFlag,
+        uint8_t bitReverseFlag);
+
+/* Deprecated */
+  void arm_cfft_radix2_q31(
+  const arm_cfft_radix2_instance_q31 * S,
+        q31_t * pSrc);
+
+  /**
+   * @brief Instance structure for the Q31 CFFT/CIFFT function.
+   */
+  typedef struct
+  {
+          uint16_t fftLen;                 /**< length of the FFT. */
+          uint8_t ifftFlag;                /**< flag that selects forward (ifftFlag=0) or inverse (ifftFlag=1) transform. */
+          uint8_t bitReverseFlag;          /**< flag that enables (bitReverseFlag=1) or disables (bitReverseFlag=0) bit reversal of output. */
+    const q31_t *pTwiddle;                 /**< points to the twiddle factor table. */
+    const uint16_t *pBitRevTable;          /**< points to the bit reversal table. */
+          uint16_t twidCoefModifier;       /**< twiddle coefficient modifier that supports different size FFTs with the same twiddle factor table. */
+          uint16_t bitRevFactor;           /**< bit reversal modifier that supports different size FFTs with the same bit reversal table. */
+  } arm_cfft_radix4_instance_q31;
+
+/* Deprecated */
+  void arm_cfft_radix4_q31(
+  const arm_cfft_radix4_instance_q31 * S,
+        q31_t * pSrc);
+
+/* Deprecated */
+  arm_status arm_cfft_radix4_init_q31(
+        arm_cfft_radix4_instance_q31 * S,
+        uint16_t fftLen,
+        uint8_t ifftFlag,
+        uint8_t bitReverseFlag);
+
+  /**
+   * @brief Instance structure for the floating-point CFFT/CIFFT function.
+   */
+  typedef struct
+  {
+          uint16_t fftLen;                   /**< length of the FFT. */
+          uint8_t ifftFlag;                  /**< flag that selects forward (ifftFlag=0) or inverse (ifftFlag=1) transform. */
+          uint8_t bitReverseFlag;            /**< flag that enables (bitReverseFlag=1) or disables (bitReverseFlag=0) bit reversal of output. */
+    const float32_t *pTwiddle;               /**< points to the Twiddle factor table. */
+    const uint16_t *pBitRevTable;            /**< points to the bit reversal table. */
+          uint16_t twidCoefModifier;         /**< twiddle coefficient modifier that supports different size FFTs with the same twiddle factor table. */
+          uint16_t bitRevFactor;             /**< bit reversal modifier that supports different size FFTs with the same bit reversal table. */
+          float32_t onebyfftLen;             /**< value of 1/fftLen. */
+  } arm_cfft_radix2_instance_f32;
+
+
+/* Deprecated */
+  arm_status arm_cfft_radix2_init_f32(
+        arm_cfft_radix2_instance_f32 * S,
+        uint16_t fftLen,
+        uint8_t ifftFlag,
+        uint8_t bitReverseFlag);
+
+/* Deprecated */
+  void arm_cfft_radix2_f32(
+  const arm_cfft_radix2_instance_f32 * S,
+        float32_t * pSrc);
+
+  /**
+   * @brief Instance structure for the floating-point CFFT/CIFFT function.
+   */
+  typedef struct
+  {
+          uint16_t fftLen;                   /**< length of the FFT. */
+          uint8_t ifftFlag;                  /**< flag that selects forward (ifftFlag=0) or inverse (ifftFlag=1) transform. */
+          uint8_t bitReverseFlag;            /**< flag that enables (bitReverseFlag=1) or disables (bitReverseFlag=0) bit reversal of output. */
+    const float32_t *pTwiddle;               /**< points to the Twiddle factor table. */
+    const uint16_t *pBitRevTable;            /**< points to the bit reversal table. */
+          uint16_t twidCoefModifier;         /**< twiddle coefficient modifier that supports different size FFTs with the same twiddle factor table. */
+          uint16_t bitRevFactor;             /**< bit reversal modifier that supports different size FFTs with the same bit reversal table. */
+          float32_t onebyfftLen;             /**< value of 1/fftLen. */
+  } arm_cfft_radix4_instance_f32;
+
+
+
+/* Deprecated */
+  arm_status arm_cfft_radix4_init_f32(
+        arm_cfft_radix4_instance_f32 * S,
+        uint16_t fftLen,
+        uint8_t ifftFlag,
+        uint8_t bitReverseFlag);
+
+/* Deprecated */
+  void arm_cfft_radix4_f32(
+  const arm_cfft_radix4_instance_f32 * S,
+        float32_t * pSrc);
+
+  /**
+   * @brief Instance structure for the fixed-point CFFT/CIFFT function.
+   */
+  typedef struct
+  {
+          uint16_t fftLen;                   /**< length of the FFT. */
+    const q15_t *pTwiddle;             /**< points to the Twiddle factor table. */
+    const uint16_t *pBitRevTable;      /**< points to the bit reversal table. */
+          uint16_t bitRevLength;             /**< bit reversal table length. */
+#if defined(ARM_MATH_MVEI) && !defined(ARM_MATH_AUTOVECTORIZE)
+   const uint32_t *rearranged_twiddle_tab_stride1_arr;        /**< Per stage reordered twiddle pointer (offset 1) */                                                       \
+   const uint32_t *rearranged_twiddle_tab_stride2_arr;        /**< Per stage reordered twiddle pointer (offset 2) */                                                       \
+   const uint32_t *rearranged_twiddle_tab_stride3_arr;        /**< Per stage reordered twiddle pointer (offset 3) */                                                       \
+   const q15_t *rearranged_twiddle_stride1; /**< reordered twiddle offset 1 storage */                                                                   \
+   const q15_t *rearranged_twiddle_stride2; /**< reordered twiddle offset 2 storage */                                                                   \
+   const q15_t *rearranged_twiddle_stride3;
+#endif
+  } arm_cfft_instance_q15;
+
+arm_status arm_cfft_init_q15(
+  arm_cfft_instance_q15 * S,
+  uint16_t fftLen);
+
+void arm_cfft_q15(
+    const arm_cfft_instance_q15 * S,
+          q15_t * p1,
+          uint8_t ifftFlag,
+          uint8_t bitReverseFlag);
+
+  /**
+   * @brief Instance structure for the fixed-point CFFT/CIFFT function.
+   */
+  typedef struct
+  {
+          uint16_t fftLen;                   /**< length of the FFT. */
+    const q31_t *pTwiddle;             /**< points to the Twiddle factor table. */
+    const uint16_t *pBitRevTable;      /**< points to the bit reversal table. */
+          uint16_t bitRevLength;             /**< bit reversal table length. */
+#if defined(ARM_MATH_MVEI) && !defined(ARM_MATH_AUTOVECTORIZE)
+   const uint32_t *rearranged_twiddle_tab_stride1_arr;        /**< Per stage reordered twiddle pointer (offset 1) */                                                       \
+   const uint32_t *rearranged_twiddle_tab_stride2_arr;        /**< Per stage reordered twiddle pointer (offset 2) */                                                       \
+   const uint32_t *rearranged_twiddle_tab_stride3_arr;        /**< Per stage reordered twiddle pointer (offset 3) */                                                       \
+   const q31_t *rearranged_twiddle_stride1; /**< reordered twiddle offset 1 storage */                                                                   \
+   const q31_t *rearranged_twiddle_stride2; /**< reordered twiddle offset 2 storage */                                                                   \
+   const q31_t *rearranged_twiddle_stride3;
+#endif
+  } arm_cfft_instance_q31;
+
+arm_status arm_cfft_init_q31(
+  arm_cfft_instance_q31 * S,
+  uint16_t fftLen);
+
+void arm_cfft_q31(
+    const arm_cfft_instance_q31 * S,
+          q31_t * p1,
+          uint8_t ifftFlag,
+          uint8_t bitReverseFlag);
+
+  /**
+   * @brief Instance structure for the floating-point CFFT/CIFFT function.
+   */
+  typedef struct
+  {
+          uint16_t fftLen;                   /**< length of the FFT. */
+    const float32_t *pTwiddle;         /**< points to the Twiddle factor table. */
+    const uint16_t *pBitRevTable;      /**< points to the bit reversal table. */
+          uint16_t bitRevLength;             /**< bit reversal table length. */
+#if defined(ARM_MATH_MVEF) && !defined(ARM_MATH_AUTOVECTORIZE)
+   const uint32_t *rearranged_twiddle_tab_stride1_arr;        /**< Per stage reordered twiddle pointer (offset 1) */                                                       \
+   const uint32_t *rearranged_twiddle_tab_stride2_arr;        /**< Per stage reordered twiddle pointer (offset 2) */                                                       \
+   const uint32_t *rearranged_twiddle_tab_stride3_arr;        /**< Per stage reordered twiddle pointer (offset 3) */                                                       \
+   const float32_t *rearranged_twiddle_stride1; /**< reordered twiddle offset 1 storage */                                                                   \
+   const float32_t *rearranged_twiddle_stride2; /**< reordered twiddle offset 2 storage */                                                                   \
+   const float32_t *rearranged_twiddle_stride3;
+#endif
+  } arm_cfft_instance_f32;
+
+
+
+  arm_status arm_cfft_init_f32(
+  arm_cfft_instance_f32 * S,
+  uint16_t fftLen);
+
+  void arm_cfft_f32(
+  const arm_cfft_instance_f32 * S,
+        float32_t * p1,
+        uint8_t ifftFlag,
+        uint8_t bitReverseFlag);
+
+
+  /**
+   * @brief Instance structure for the Double Precision Floating-point CFFT/CIFFT function.
+   */
+  typedef struct
+  {
+          uint16_t fftLen;                   /**< length of the FFT. */
+    const float64_t *pTwiddle;         /**< points to the Twiddle factor table. */
+    const uint16_t *pBitRevTable;      /**< points to the bit reversal table. */
+          uint16_t bitRevLength;             /**< bit reversal table length. */
+  } arm_cfft_instance_f64;
+
+  arm_status arm_cfft_init_f64(
+  arm_cfft_instance_f64 * S,
+  uint16_t fftLen);
+  
+  void arm_cfft_f64(
+  const arm_cfft_instance_f64 * S,
+        float64_t * p1,
+        uint8_t ifftFlag,
+        uint8_t bitReverseFlag);
+
+  /**
+   * @brief Instance structure for the Q15 RFFT/RIFFT function.
+   */
+  typedef struct
+  {
+          uint32_t fftLenReal;                      /**< length of the real FFT. */
+          uint8_t ifftFlagR;                        /**< flag that selects forward (ifftFlagR=0) or inverse (ifftFlagR=1) transform. */
+          uint8_t bitReverseFlagR;                  /**< flag that enables (bitReverseFlagR=1) or disables (bitReverseFlagR=0) bit reversal of output. */
+          uint32_t twidCoefRModifier;               /**< twiddle coefficient modifier that supports different size FFTs with the same twiddle factor table. */
+    const q15_t *pTwiddleAReal;                     /**< points to the real twiddle factor table. */
+    const q15_t *pTwiddleBReal;                     /**< points to the imag twiddle factor table. */
+#if defined(ARM_MATH_MVEI) && !defined(ARM_MATH_AUTOVECTORIZE)
+    arm_cfft_instance_q15 cfftInst;
+#else
+    const arm_cfft_instance_q15 *pCfft;       /**< points to the complex FFT instance. */
+#endif
+  } arm_rfft_instance_q15;
+
+  arm_status arm_rfft_init_q15(
+        arm_rfft_instance_q15 * S,
+        uint32_t fftLenReal,
+        uint32_t ifftFlagR,
+        uint32_t bitReverseFlag);
+
+  void arm_rfft_q15(
+  const arm_rfft_instance_q15 * S,
+        q15_t * pSrc,
+        q15_t * pDst);
+
+  /**
+   * @brief Instance structure for the Q31 RFFT/RIFFT function.
+   */
+  typedef struct
+  {
+          uint32_t fftLenReal;                        /**< length of the real FFT. */
+          uint8_t ifftFlagR;                          /**< flag that selects forward (ifftFlagR=0) or inverse (ifftFlagR=1) transform. */
+          uint8_t bitReverseFlagR;                    /**< flag that enables (bitReverseFlagR=1) or disables (bitReverseFlagR=0) bit reversal of output. */
+          uint32_t twidCoefRModifier;                 /**< twiddle coefficient modifier that supports different size FFTs with the same twiddle factor table. */
+    const q31_t *pTwiddleAReal;                       /**< points to the real twiddle factor table. */
+    const q31_t *pTwiddleBReal;                       /**< points to the imag twiddle factor table. */
+#if defined(ARM_MATH_MVEI) && !defined(ARM_MATH_AUTOVECTORIZE)
+    arm_cfft_instance_q31 cfftInst;
+#else
+    const arm_cfft_instance_q31 *pCfft;         /**< points to the complex FFT instance. */
+#endif
+  } arm_rfft_instance_q31;
+
+  arm_status arm_rfft_init_q31(
+        arm_rfft_instance_q31 * S,
+        uint32_t fftLenReal,
+        uint32_t ifftFlagR,
+        uint32_t bitReverseFlag);
+
+  void arm_rfft_q31(
+  const arm_rfft_instance_q31 * S,
+        q31_t * pSrc,
+        q31_t * pDst);
+
+  /**
+   * @brief Instance structure for the floating-point RFFT/RIFFT function.
+   */
+  typedef struct
+  {
+          uint32_t fftLenReal;                        /**< length of the real FFT. */
+          uint16_t fftLenBy2;                         /**< length of the complex FFT. */
+          uint8_t ifftFlagR;                          /**< flag that selects forward (ifftFlagR=0) or inverse (ifftFlagR=1) transform. */
+          uint8_t bitReverseFlagR;                    /**< flag that enables (bitReverseFlagR=1) or disables (bitReverseFlagR=0) bit reversal of output. */
+          uint32_t twidCoefRModifier;                     /**< twiddle coefficient modifier that supports different size FFTs with the same twiddle factor table. */
+    const float32_t *pTwiddleAReal;                   /**< points to the real twiddle factor table. */
+    const float32_t *pTwiddleBReal;                   /**< points to the imag twiddle factor table. */
+          arm_cfft_radix4_instance_f32 *pCfft;        /**< points to the complex FFT instance. */
+  } arm_rfft_instance_f32;
+
+  arm_status arm_rfft_init_f32(
+        arm_rfft_instance_f32 * S,
+        arm_cfft_radix4_instance_f32 * S_CFFT,
+        uint32_t fftLenReal,
+        uint32_t ifftFlagR,
+        uint32_t bitReverseFlag);
+
+  void arm_rfft_f32(
+  const arm_rfft_instance_f32 * S,
+        float32_t * pSrc,
+        float32_t * pDst);
+
+  /**
+   * @brief Instance structure for the Double Precision Floating-point RFFT/RIFFT function.
+   */
+typedef struct
+  {
+          arm_cfft_instance_f64 Sint;      /**< Internal CFFT structure. */
+          uint16_t fftLenRFFT;             /**< length of the real sequence */
+    const float64_t * pTwiddleRFFT;        /**< Twiddle factors real stage  */
+  } arm_rfft_fast_instance_f64 ;
+
+arm_status arm_rfft_fast_init_f64 (
+         arm_rfft_fast_instance_f64 * S,
+         uint16_t fftLen);
+
+
+void arm_rfft_fast_f64(
+    arm_rfft_fast_instance_f64 * S,
+    float64_t * p, float64_t * pOut,
+    uint8_t ifftFlag);
+
+
+  /**
+   * @brief Instance structure for the floating-point RFFT/RIFFT function.
+   */
+typedef struct
+  {
+          arm_cfft_instance_f32 Sint;      /**< Internal CFFT structure. */
+          uint16_t fftLenRFFT;             /**< length of the real sequence */
+    const float32_t * pTwiddleRFFT;        /**< Twiddle factors real stage  */
+  } arm_rfft_fast_instance_f32 ;
+
+arm_status arm_rfft_fast_init_f32 (
+         arm_rfft_fast_instance_f32 * S,
+         uint16_t fftLen);
+
+
+  void arm_rfft_fast_f32(
+        const arm_rfft_fast_instance_f32 * S,
+        float32_t * p, float32_t * pOut,
+        uint8_t ifftFlag);
+
+  /**
+   * @brief Instance structure for the floating-point DCT4/IDCT4 function.
+   */
+  typedef struct
+  {
+          uint16_t N;                          /**< length of the DCT4. */
+          uint16_t Nby2;                       /**< half of the length of the DCT4. */
+          float32_t normalize;                 /**< normalizing factor. */
+    const float32_t *pTwiddle;                 /**< points to the twiddle factor table. */
+    const float32_t *pCosFactor;               /**< points to the cosFactor table. */
+          arm_rfft_instance_f32 *pRfft;        /**< points to the real FFT instance. */
+          arm_cfft_radix4_instance_f32 *pCfft; /**< points to the complex FFT instance. */
+  } arm_dct4_instance_f32;
+
+
+  /**
+   * @brief  Initialization function for the floating-point DCT4/IDCT4.
+   * @param[in,out] S          points to an instance of floating-point DCT4/IDCT4 structure.
+   * @param[in]     S_RFFT     points to an instance of floating-point RFFT/RIFFT structure.
+   * @param[in]     S_CFFT     points to an instance of floating-point CFFT/CIFFT structure.
+   * @param[in]     N          length of the DCT4.
+   * @param[in]     Nby2       half of the length of the DCT4.
+   * @param[in]     normalize  normalizing factor.
+   * @return      arm_status function returns ARM_MATH_SUCCESS if initialization is successful or ARM_MATH_ARGUMENT_ERROR if <code>fftLenReal</code> is not a supported transform length.
+   */
+  arm_status arm_dct4_init_f32(
+        arm_dct4_instance_f32 * S,
+        arm_rfft_instance_f32 * S_RFFT,
+        arm_cfft_radix4_instance_f32 * S_CFFT,
+        uint16_t N,
+        uint16_t Nby2,
+        float32_t normalize);
+
+
+  /**
+   * @brief Processing function for the floating-point DCT4/IDCT4.
+   * @param[in]     S              points to an instance of the floating-point DCT4/IDCT4 structure.
+   * @param[in]     pState         points to state buffer.
+   * @param[in,out] pInlineBuffer  points to the in-place input and output buffer.
+   */
+  void arm_dct4_f32(
+  const arm_dct4_instance_f32 * S,
+        float32_t * pState,
+        float32_t * pInlineBuffer);
+
+
+  /**
+   * @brief Instance structure for the Q31 DCT4/IDCT4 function.
+   */
+  typedef struct
+  {
+          uint16_t N;                          /**< length of the DCT4. */
+          uint16_t Nby2;                       /**< half of the length of the DCT4. */
+          q31_t normalize;                     /**< normalizing factor. */
+    const q31_t *pTwiddle;                     /**< points to the twiddle factor table. */
+    const q31_t *pCosFactor;                   /**< points to the cosFactor table. */
+          arm_rfft_instance_q31 *pRfft;        /**< points to the real FFT instance. */
+          arm_cfft_radix4_instance_q31 *pCfft; /**< points to the complex FFT instance. */
+  } arm_dct4_instance_q31;
+
+
+  /**
+   * @brief  Initialization function for the Q31 DCT4/IDCT4.
+   * @param[in,out] S          points to an instance of Q31 DCT4/IDCT4 structure.
+   * @param[in]     S_RFFT     points to an instance of Q31 RFFT/RIFFT structure
+   * @param[in]     S_CFFT     points to an instance of Q31 CFFT/CIFFT structure
+   * @param[in]     N          length of the DCT4.
+   * @param[in]     Nby2       half of the length of the DCT4.
+   * @param[in]     normalize  normalizing factor.
+   * @return      arm_status function returns ARM_MATH_SUCCESS if initialization is successful or ARM_MATH_ARGUMENT_ERROR if <code>N</code> is not a supported transform length.
+   */
+  arm_status arm_dct4_init_q31(
+        arm_dct4_instance_q31 * S,
+        arm_rfft_instance_q31 * S_RFFT,
+        arm_cfft_radix4_instance_q31 * S_CFFT,
+        uint16_t N,
+        uint16_t Nby2,
+        q31_t normalize);
+
+
+  /**
+   * @brief Processing function for the Q31 DCT4/IDCT4.
+   * @param[in]     S              points to an instance of the Q31 DCT4 structure.
+   * @param[in]     pState         points to state buffer.
+   * @param[in,out] pInlineBuffer  points to the in-place input and output buffer.
+   */
+  void arm_dct4_q31(
+  const arm_dct4_instance_q31 * S,
+        q31_t * pState,
+        q31_t * pInlineBuffer);
+
+
+  /**
+   * @brief Instance structure for the Q15 DCT4/IDCT4 function.
+   */
+  typedef struct
+  {
+          uint16_t N;                          /**< length of the DCT4. */
+          uint16_t Nby2;                       /**< half of the length of the DCT4. */
+          q15_t normalize;                     /**< normalizing factor. */
+    const q15_t *pTwiddle;                     /**< points to the twiddle factor table. */
+    const q15_t *pCosFactor;                   /**< points to the cosFactor table. */
+          arm_rfft_instance_q15 *pRfft;        /**< points to the real FFT instance. */
+          arm_cfft_radix4_instance_q15 *pCfft; /**< points to the complex FFT instance. */
+  } arm_dct4_instance_q15;
+
+
+  /**
+   * @brief  Initialization function for the Q15 DCT4/IDCT4.
+   * @param[in,out] S          points to an instance of Q15 DCT4/IDCT4 structure.
+   * @param[in]     S_RFFT     points to an instance of Q15 RFFT/RIFFT structure.
+   * @param[in]     S_CFFT     points to an instance of Q15 CFFT/CIFFT structure.
+   * @param[in]     N          length of the DCT4.
+   * @param[in]     Nby2       half of the length of the DCT4.
+   * @param[in]     normalize  normalizing factor.
+   * @return      arm_status function returns ARM_MATH_SUCCESS if initialization is successful or ARM_MATH_ARGUMENT_ERROR if <code>N</code> is not a supported transform length.
+   */
+  arm_status arm_dct4_init_q15(
+        arm_dct4_instance_q15 * S,
+        arm_rfft_instance_q15 * S_RFFT,
+        arm_cfft_radix4_instance_q15 * S_CFFT,
+        uint16_t N,
+        uint16_t Nby2,
+        q15_t normalize);
+
+
+  /**
+   * @brief Processing function for the Q15 DCT4/IDCT4.
+   * @param[in]     S              points to an instance of the Q15 DCT4 structure.
+   * @param[in]     pState         points to state buffer.
+   * @param[in,out] pInlineBuffer  points to the in-place input and output buffer.
+   */
+  void arm_dct4_q15(
+  const arm_dct4_instance_q15 * S,
+        q15_t * pState,
+        q15_t * pInlineBuffer);
+
+  /**
+   * @brief Instance structure for the Floating-point MFCC function.
+   */
+typedef struct
+  {
+     const float32_t *dctCoefs; /**< Internal DCT coefficients */
+     const float32_t *filterCoefs; /**< Internal Mel filter coefficients */ 
+     const float32_t *windowCoefs; /**< Windowing coefficients */ 
+     const uint32_t *filterPos; /**< Internal Mel filter positions in spectrum */ 
+     const uint32_t *filterLengths; /**< Internal Mel filter  lengths */ 
+     uint32_t fftLen; /**< FFT length */
+     uint32_t nbMelFilters; /**< Number of Mel filters */
+     uint32_t nbDctOutputs; /**< Number of DCT outputs */
+#if defined(ARM_MFCC_CFFT_BASED)
+     /* Implementation of the MFCC is using a CFFT */
+     arm_cfft_instance_f32 cfft; /**< Internal CFFT instance */
+#else
+     /* Implementation of the MFCC is using a RFFT (default) */
+     arm_rfft_fast_instance_f32 rfft;
+#endif
+  } arm_mfcc_instance_f32 ;
+
+arm_status arm_mfcc_init_f32(
+  arm_mfcc_instance_f32 * S,
+  uint32_t fftLen,
+  uint32_t nbMelFilters,
+  uint32_t nbDctOutputs,
+  const float32_t *dctCoefs,
+  const uint32_t *filterPos,
+  const uint32_t *filterLengths,
+  const float32_t *filterCoefs,
+  const float32_t *windowCoefs
+  );
+
+
+/**
+  @brief         MFCC F32
+  @param[in]    S       points to the mfcc instance structure
+  @param[in]     pSrc points to the input samples
+  @param[out]     pDst  points to the output MFCC values
+  @param[inout]     pTmp  points to a temporary buffer of complex
+  @return        none
+ */
+  void arm_mfcc_f32(
+  const arm_mfcc_instance_f32 * S,
+  float32_t *pSrc,
+  float32_t *pDst,
+  float32_t *pTmp
+  );
+
+typedef struct
+  {
+     const q31_t *dctCoefs; /**< Internal DCT coefficients */
+     const q31_t *filterCoefs; /**< Internal Mel filter coefficients */ 
+     const q31_t *windowCoefs; /**< Windowing coefficients */ 
+     const uint32_t *filterPos; /**< Internal Mel filter positions in spectrum */ 
+     const uint32_t *filterLengths; /**< Internal Mel filter  lengths */ 
+     uint32_t fftLen; /**< FFT length */
+     uint32_t nbMelFilters; /**< Number of Mel filters */
+     uint32_t nbDctOutputs; /**< Number of DCT outputs */
+#if defined(ARM_MFCC_CFFT_BASED)
+     /* Implementation of the MFCC is using a CFFT */
+     arm_cfft_instance_q31 cfft; /**< Internal CFFT instance */
+#else
+     /* Implementation of the MFCC is using a RFFT (default) */
+     arm_rfft_instance_q31 rfft;
+#endif
+  } arm_mfcc_instance_q31 ;
+
+arm_status arm_mfcc_init_q31(
+  arm_mfcc_instance_q31 * S,
+  uint32_t fftLen,
+  uint32_t nbMelFilters,
+  uint32_t nbDctOutputs,
+  const q31_t *dctCoefs,
+  const uint32_t *filterPos,
+  const uint32_t *filterLengths,
+  const q31_t *filterCoefs,
+  const q31_t *windowCoefs
+  );
+
+
+/**
+  @brief         MFCC Q31
+  @param[in]    S       points to the mfcc instance structure
+  @param[in]     pSrc points to the input samples
+  @param[out]     pDst  points to the output MFCC values
+  @param[inout]     pTmp  points to a temporary buffer of complex
+  @return        none
+ */
+  arm_status arm_mfcc_q31(
+  const arm_mfcc_instance_q31 * S,
+  q31_t *pSrc,
+  q31_t *pDst,
+  q31_t *pTmp
+  );
+
+typedef struct
+  {
+     const q15_t *dctCoefs; /**< Internal DCT coefficients */
+     const q15_t *filterCoefs; /**< Internal Mel filter coefficients */ 
+     const q15_t *windowCoefs; /**< Windowing coefficients */ 
+     const uint32_t *filterPos; /**< Internal Mel filter positions in spectrum */ 
+     const uint32_t *filterLengths; /**< Internal Mel filter  lengths */ 
+     uint32_t fftLen; /**< FFT length */
+     uint32_t nbMelFilters; /**< Number of Mel filters */
+     uint32_t nbDctOutputs; /**< Number of DCT outputs */
+#if defined(ARM_MFCC_CFFT_BASED)
+     /* Implementation of the MFCC is using a CFFT */
+     arm_cfft_instance_q15 cfft; /**< Internal CFFT instance */
+#else
+     /* Implementation of the MFCC is using a RFFT (default) */
+     arm_rfft_instance_q15 rfft;
+#endif
+  } arm_mfcc_instance_q15 ;
+
+arm_status arm_mfcc_init_q15(
+  arm_mfcc_instance_q15 * S,
+  uint32_t fftLen,
+  uint32_t nbMelFilters,
+  uint32_t nbDctOutputs,
+  const q15_t *dctCoefs,
+  const uint32_t *filterPos,
+  const uint32_t *filterLengths,
+  const q15_t *filterCoefs,
+  const q15_t *windowCoefs
+  );
+
+
+/**
+  @brief         MFCC Q15
+  @param[in]    S       points to the mfcc instance structure
+  @param[in]     pSrc points to the input samples
+  @param[out]     pDst  points to the output MFCC values in q8.7 format
+  @param[inout]     pTmp  points to a temporary buffer of complex
+  @return        error status
+ */
+  arm_status arm_mfcc_q15(
+  const arm_mfcc_instance_q15 * S,
+  q15_t *pSrc,
+  q15_t *pDst,
+  q31_t *pTmp
+  );
+
+
+#ifdef   __cplusplus
+}
+#endif
+
+#endif /* ifndef _TRANSFORM_FUNCTIONS_H_ */

bsp/cmsis/include/dsp/transform_functions_f16.h

@@ -0,0 +1,208 @@
+/******************************************************************************
+ * @file     transform_functions_f16.h
+ * @brief    Public header file for CMSIS DSP Library
+ * @version  V1.10.0
+ * @date     08 July 2021
+ * Target Processor: Cortex-M and Cortex-A cores
+ ******************************************************************************/
+/*
+ * Copyright (c) 2010-2020 Arm Limited or its affiliates. All rights reserved.
+ *
+ * SPDX-License-Identifier: Apache-2.0
+ *
+ * Licensed under the Apache License, Version 2.0 (the License); you may
+ * not use this file except in compliance with the License.
+ * You may obtain a copy of the License at
+ *
+ * www.apache.org/licenses/LICENSE-2.0
+ *
+ * Unless required by applicable law or agreed to in writing, software
+ * distributed under the License is distributed on an AS IS BASIS, WITHOUT
+ * WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ * See the License for the specific language governing permissions and
+ * limitations under the License.
+ */
+
+ 
+#ifndef _TRANSFORM_FUNCTIONS_F16_H_
+#define _TRANSFORM_FUNCTIONS_F16_H_
+
+#include "arm_math_types_f16.h"
+#include "arm_math_memory.h"
+
+#include "dsp/none.h"
+#include "dsp/utils.h"
+
+#ifdef   __cplusplus
+extern "C"
+{
+#endif
+
+
+
+#if defined(ARM_FLOAT16_SUPPORTED)
+
+
+  /**
+   * @brief Instance structure for the floating-point CFFT/CIFFT function.
+   */
+  typedef struct
+  {
+          uint16_t fftLen;                   /**< length of the FFT. */
+          uint8_t ifftFlag;                  /**< flag that selects forward (ifftFlag=0) or inverse (ifftFlag=1) transform. */
+          uint8_t bitReverseFlag;            /**< flag that enables (bitReverseFlag=1) or disables (bitReverseFlag=0) bit reversal of output. */
+    const float16_t *pTwiddle;               /**< points to the Twiddle factor table. */
+    const uint16_t *pBitRevTable;            /**< points to the bit reversal table. */
+          uint16_t twidCoefModifier;         /**< twiddle coefficient modifier that supports different size FFTs with the same twiddle factor table. */
+          uint16_t bitRevFactor;             /**< bit reversal modifier that supports different size FFTs with the same bit reversal table. */
+          float16_t onebyfftLen;             /**< value of 1/fftLen. */
+  } arm_cfft_radix2_instance_f16;
+
+  /**
+   * @brief Instance structure for the floating-point CFFT/CIFFT function.
+   */
+  typedef struct
+  {
+          uint16_t fftLen;                   /**< length of the FFT. */
+          uint8_t ifftFlag;                  /**< flag that selects forward (ifftFlag=0) or inverse (ifftFlag=1) transform. */
+          uint8_t bitReverseFlag;            /**< flag that enables (bitReverseFlag=1) or disables (bitReverseFlag=0) bit reversal of output. */
+    const float16_t *pTwiddle;               /**< points to the Twiddle factor table. */
+    const uint16_t *pBitRevTable;            /**< points to the bit reversal table. */
+          uint16_t twidCoefModifier;         /**< twiddle coefficient modifier that supports different size FFTs with the same twiddle factor table. */
+          uint16_t bitRevFactor;             /**< bit reversal modifier that supports different size FFTs with the same bit reversal table. */
+          float16_t onebyfftLen;             /**< value of 1/fftLen. */
+  } arm_cfft_radix4_instance_f16;
+
+  /**
+   * @brief Instance structure for the floating-point CFFT/CIFFT function.
+   */
+  typedef struct
+  {
+          uint16_t fftLen;                   /**< length of the FFT. */
+    const float16_t *pTwiddle;         /**< points to the Twiddle factor table. */
+    const uint16_t *pBitRevTable;      /**< points to the bit reversal table. */
+          uint16_t bitRevLength;             /**< bit reversal table length. */
+#if defined(ARM_MATH_MVEF) && !defined(ARM_MATH_AUTOVECTORIZE)
+   const uint32_t *rearranged_twiddle_tab_stride1_arr;        /**< Per stage reordered twiddle pointer (offset 1) */                                                       \
+   const uint32_t *rearranged_twiddle_tab_stride2_arr;        /**< Per stage reordered twiddle pointer (offset 2) */                                                       \
+   const uint32_t *rearranged_twiddle_tab_stride3_arr;        /**< Per stage reordered twiddle pointer (offset 3) */                                                       \
+   const float16_t *rearranged_twiddle_stride1; /**< reordered twiddle offset 1 storage */                                                                   \
+   const float16_t *rearranged_twiddle_stride2; /**< reordered twiddle offset 2 storage */                                                                   \
+   const float16_t *rearranged_twiddle_stride3;
+#endif
+  } arm_cfft_instance_f16;
+
+
+  arm_status arm_cfft_init_f16(
+  arm_cfft_instance_f16 * S,
+  uint16_t fftLen);
+
+  void arm_cfft_f16(
+  const arm_cfft_instance_f16 * S,
+        float16_t * p1,
+        uint8_t ifftFlag,
+        uint8_t bitReverseFlag);
+
+  /**
+   * @brief Instance structure for the floating-point RFFT/RIFFT function.
+   */
+typedef struct
+  {
+          arm_cfft_instance_f16 Sint;      /**< Internal CFFT structure. */
+          uint16_t fftLenRFFT;             /**< length of the real sequence */
+    const float16_t * pTwiddleRFFT;        /**< Twiddle factors real stage  */
+  } arm_rfft_fast_instance_f16 ;
+
+arm_status arm_rfft_fast_init_f16 (
+         arm_rfft_fast_instance_f16 * S,
+         uint16_t fftLen);
+
+
+  void arm_rfft_fast_f16(
+        const arm_rfft_fast_instance_f16 * S,
+        float16_t * p, float16_t * pOut,
+        uint8_t ifftFlag);
+
+/* Deprecated */
+  arm_status arm_cfft_radix4_init_f16(
+        arm_cfft_radix4_instance_f16 * S,
+        uint16_t fftLen,
+        uint8_t ifftFlag,
+        uint8_t bitReverseFlag);
+
+/* Deprecated */
+  void arm_cfft_radix4_f16(
+  const arm_cfft_radix4_instance_f16 * S,
+        float16_t * pSrc);
+
+
+/* Deprecated */
+  arm_status arm_cfft_radix2_init_f16(
+        arm_cfft_radix2_instance_f16 * S,
+        uint16_t fftLen,
+        uint8_t ifftFlag,
+        uint8_t bitReverseFlag);
+
+/* Deprecated */
+  void arm_cfft_radix2_f16(
+  const arm_cfft_radix2_instance_f16 * S,
+        float16_t * pSrc);
+
+  /**
+   * @brief Instance structure for the Floating-point MFCC function.
+   */
+typedef struct
+  {
+     const float16_t *dctCoefs; /**< Internal DCT coefficients */
+     const float16_t *filterCoefs; /**< Internal Mel filter coefficients */ 
+     const float16_t *windowCoefs; /**< Windowing coefficients */ 
+     const uint32_t *filterPos; /**< Internal Mel filter positions in spectrum */ 
+     const uint32_t *filterLengths; /**< Internal Mel filter  lengths */ 
+     uint32_t fftLen; /**< FFT length */
+     uint32_t nbMelFilters; /**< Number of Mel filters */
+     uint32_t nbDctOutputs; /**< Number of DCT outputs */
+#if defined(ARM_MFCC_CFFT_BASED)
+     /* Implementation of the MFCC is using a CFFT */
+     arm_cfft_instance_f16 cfft; /**< Internal CFFT instance */
+#else
+     /* Implementation of the MFCC is using a RFFT (default) */
+     arm_rfft_fast_instance_f16 rfft;
+#endif
+  } arm_mfcc_instance_f16 ;
+
+arm_status arm_mfcc_init_f16(
+  arm_mfcc_instance_f16 * S,
+  uint32_t fftLen,
+  uint32_t nbMelFilters,
+  uint32_t nbDctOutputs,
+  const float16_t *dctCoefs,
+  const uint32_t *filterPos,
+  const uint32_t *filterLengths,
+  const float16_t *filterCoefs,
+  const float16_t *windowCoefs
+  );
+
+
+/**
+  @brief         MFCC F16
+  @param[in]    S       points to the mfcc instance structure
+  @param[in]     pSrc points to the input samples
+  @param[out]     pDst  points to the output MFCC values
+  @param[inout]     pTmp  points to a temporary buffer of complex
+  @return        none
+ */
+  void arm_mfcc_f16(
+  const arm_mfcc_instance_f16 * S,
+  float16_t *pSrc,
+  float16_t *pDst,
+  float16_t *pTmp
+  );
+
+  
+#endif /* defined(ARM_FLOAT16_SUPPORTED)*/
+
+#ifdef   __cplusplus
+}
+#endif
+
+#endif /* ifndef _TRANSFORM_FUNCTIONS_F16_H_ */

bsp/cmsis/include/dsp/utils.h

@@ -0,0 +1,240 @@
+/******************************************************************************
+ * @file     arm_math_utils.h
+ * @brief    Public header file for CMSIS DSP Library
+ * @version  V1.9.0
+ * @date     20. July 2020
+ ******************************************************************************/
+/*
+ * Copyright (c) 2010-2020 Arm Limited or its affiliates. All rights reserved.
+ *
+ * SPDX-License-Identifier: Apache-2.0
+ *
+ * Licensed under the Apache License, Version 2.0 (the License); you may
+ * not use this file except in compliance with the License.
+ * You may obtain a copy of the License at
+ *
+ * www.apache.org/licenses/LICENSE-2.0
+ *
+ * Unless required by applicable law or agreed to in writing, software
+ * distributed under the License is distributed on an AS IS BASIS, WITHOUT
+ * WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ * See the License for the specific language governing permissions and
+ * limitations under the License.
+ */
+
+#ifndef _ARM_MATH_UTILS_H_
+
+#define _ARM_MATH_UTILS_H_
+
+#include "arm_math_types.h"
+
+#ifdef   __cplusplus
+extern "C"
+{
+#endif
+
+  /**
+   * @brief Macros required for reciprocal calculation in Normalized LMS
+   */
+
+#define INDEX_MASK         0x0000003F
+
+
+#define SQ(x) ((x) * (x))
+
+#define ROUND_UP(N, S) ((((N) + (S) - 1) / (S)) * (S))
+
+
+  /**
+   * @brief Function to Calculates 1/in (reciprocal) value of Q31 Data type.
+   */
+  __STATIC_FORCEINLINE uint32_t arm_recip_q31(
+        q31_t in,
+        q31_t * dst,
+  const q31_t * pRecipTable)
+  {
+    q31_t out;
+    uint32_t tempVal;
+    uint32_t index, i;
+    uint32_t signBits;
+
+    if (in > 0)
+    {
+      signBits = ((uint32_t) (__CLZ( in) - 1));
+    }
+    else
+    {
+      signBits = ((uint32_t) (__CLZ(-in) - 1));
+    }
+
+    /* Convert input sample to 1.31 format */
+    in = (in << signBits);
+
+    /* calculation of index for initial approximated Val */
+    index = (uint32_t)(in >> 24);
+    index = (index & INDEX_MASK);
+
+    /* 1.31 with exp 1 */
+    out = pRecipTable[index];
+
+    /* calculation of reciprocal value */
+    /* running approximation for two iterations */
+    for (i = 0U; i < 2U; i++)
+    {
+      tempVal = (uint32_t) (((q63_t) in * out) >> 31);
+      tempVal = 0x7FFFFFFFu - tempVal;
+      /*      1.31 with exp 1 */
+      /* out = (q31_t) (((q63_t) out * tempVal) >> 30); */
+      out = clip_q63_to_q31(((q63_t) out * tempVal) >> 30);
+    }
+
+    /* write output */
+    *dst = out;
+
+    /* return num of signbits of out = 1/in value */
+    return (signBits + 1U);
+  }
+
+
+  /**
+   * @brief Function to Calculates 1/in (reciprocal) value of Q15 Data type.
+   */
+  __STATIC_FORCEINLINE uint32_t arm_recip_q15(
+        q15_t in,
+        q15_t * dst,
+  const q15_t * pRecipTable)
+  {
+    q15_t out = 0;
+    uint32_t tempVal = 0;
+    uint32_t index = 0, i = 0;
+    uint32_t signBits = 0;
+
+    if (in > 0)
+    {
+      signBits = ((uint32_t)(__CLZ( in) - 17));
+    }
+    else
+    {
+      signBits = ((uint32_t)(__CLZ(-in) - 17));
+    }
+
+    /* Convert input sample to 1.15 format */
+    in = (in << signBits);
+
+    /* calculation of index for initial approximated Val */
+    index = (uint32_t)(in >>  8);
+    index = (index & INDEX_MASK);
+
+    /*      1.15 with exp 1  */
+    out = pRecipTable[index];
+
+    /* calculation of reciprocal value */
+    /* running approximation for two iterations */
+    for (i = 0U; i < 2U; i++)
+    {
+      tempVal = (uint32_t) (((q31_t) in * out) >> 15);
+      tempVal = 0x7FFFu - tempVal;
+      /*      1.15 with exp 1 */
+      out = (q15_t) (((q31_t) out * tempVal) >> 14);
+      /* out = clip_q31_to_q15(((q31_t) out * tempVal) >> 14); */
+    }
+
+    /* write output */
+    *dst = out;
+
+    /* return num of signbits of out = 1/in value */
+    return (signBits + 1);
+  }
+
+
+/**
+ * @brief  64-bit to 32-bit unsigned normalization
+ * @param[in]  in           is input unsigned long long value
+ * @param[out] normalized   is the 32-bit normalized value
+ * @param[out] norm         is norm scale
+ */
+__STATIC_INLINE  void arm_norm_64_to_32u(uint64_t in, int32_t * normalized, int32_t *norm)
+{
+    int32_t     n1;
+    int32_t     hi = (int32_t) (in >> 32);
+    int32_t     lo = (int32_t) ((in << 32) >> 32);
+
+    n1 = __CLZ(hi) - 32;
+    if (!n1)
+    {
+        /*
+         * input fits in 32-bit
+         */
+        n1 = __CLZ(lo);
+        if (!n1)
+        {
+            /*
+             * MSB set, need to scale down by 1
+             */
+            *norm = -1;
+            *normalized = (((uint32_t) lo) >> 1);
+        } else
+        {
+            if (n1 == 32)
+            {
+                /*
+                 * input is zero
+                 */
+                *norm = 0;
+                *normalized = 0;
+            } else
+            {
+                /*
+                 * 32-bit normalization
+                 */
+                *norm = n1 - 1;
+                *normalized = lo << *norm;
+            }
+        }
+    } else
+    {
+        /*
+         * input fits in 64-bit
+         */
+        n1 = 1 - n1;
+        *norm = -n1;
+        /*
+         * 64 bit normalization
+         */
+        *normalized = (((uint32_t) lo) >> n1) | (hi << (32 - n1));
+    }
+}
+
+__STATIC_INLINE q31_t arm_div_q63_to_q31(q63_t num, q31_t den)
+{
+    q31_t   result;
+    uint64_t   absNum;
+    int32_t   normalized;
+    int32_t   norm;
+
+    /*
+     * if sum fits in 32bits
+     * avoid costly 64-bit division
+     */
+    absNum = num > 0 ? num : -num;
+    arm_norm_64_to_32u(absNum, &normalized, &norm);
+    if (norm > 0)
+        /*
+         * 32-bit division
+         */
+        result = (q31_t) num / den;
+    else
+        /*
+         * 64-bit division
+         */
+        result = (q31_t) (num / den);
+
+    return result;
+}
+
+
+#ifdef   __cplusplus
+}
+#endif
+
+#endif /*ifndef _ARM_MATH_UTILS_H_ */

bsp/common/include/bsp_common.h

@@ -327,11 +327,6 @@ uint8_t OS_IsSchedulerRun(void);
 #endif
 uint32_t OS_EnterCritical(void);
 void OS_ExitCritical(uint32_t Critical);
-void *OS_Malloc(uint32_t Size);
-void *OS_Calloc(uint32_t count, uint32_t eltsize);
-void *OS_Zalloc(uint32_t Size);
-void OS_Free(void *p);
-void *OS_Realloc(void *buf, uint32_t size);
 void OS_MemInfo(uint32_t *curalloc, uint32_t *totfree, uint32_t *maxfree);
 int32_t OS_InitBuffer(Buffer_Struct *Buf, uint32_t Size);
 void OS_DeInitBuffer(Buffer_Struct *Buf);
@@ -549,11 +544,11 @@ void BytesPutDoubleToBuf(Buffer_Struct *Buf, double v);
 
 extern uint64_t GetSysTickMS();
 
-#define malloc OS_Malloc
-#define free OS_Free
-#define realloc OS_Realloc
-#define zalloc OS_Zalloc
-#define calloc OS_Calloc
+//void *__wrap_malloc(size_t Size);
+//void *__wrap_zalloc(size_t Size);
+//void *__wrap_calloc(size_t count, size_t eltsize);
+//void *__wrap_realloc(void *buf, size_t size);
+//void __wrap_free(void *p);
 
 #ifndef ASSERT
 #if defined(__DEBUG__)

bsp/common/src/bsp_common.c

@@ -20,6 +20,7 @@
  */
 
 #include "bsp_common.h"
+#include "time.h"
 #ifdef __LUATOS__
 
 
@@ -182,56 +183,6 @@ __attribute__((weak)) void OS_ExitCritical(uint32_t Critical)
 	}
 }
 
-__attribute__((weak)) void *OS_Malloc(uint32_t Size)
-{
-	void *p;
-	uint32_t Critical = OS_EnterCritical();
-	p = bget(Size);
-	OS_ExitCritical(Critical);
-	return p;
-}
-
-__attribute__((weak)) void *OS_Zalloc(uint32_t Size)
-{
-	void *p;
-	uint32_t Critical = OS_EnterCritical();
-	p = bgetz(Size);
-	OS_ExitCritical(Critical);
-	return p;
-}
-
-__attribute__((weak)) void *OS_Calloc(uint32_t count, uint32_t eltsize)
-{
-	void *p;
-	uint32_t Critical = OS_EnterCritical();
-	p = bgetz(count * eltsize);
-	OS_ExitCritical(Critical);
-	return p;
-}
-
-__attribute__((weak)) void OS_Free(void *p)
-{
-	if (((uint32_t)p >= (uint32_t)(&__os_heap_start)) && ((uint32_t)p <= (uint32_t)(&__ram_end)))
-	{
-		uint32_t Critical = OS_EnterCritical();
-		brel(p);
-		OS_ExitCritical(Critical);
-	}
-
-}
-
-__attribute__((weak)) void *OS_Realloc(void *buf, uint32_t size)
-{
-	void *p;
-	uint32_t Critical = OS_EnterCritical();
-//	p = bget(size);
-//	memcpy(p, buf, size);
-//	brel(buf);
-	p = bgetr(buf, size);
-	OS_ExitCritical(Critical);
-	return p;
-}
-
 __attribute__((weak)) void OS_MemInfo(uint32_t *curalloc, uint32_t *totfree, uint32_t *maxfree)
 {
 	unsigned long  nget, nrel;
@@ -244,7 +195,7 @@ int32_t OS_InitBuffer(Buffer_Struct *Buf, uint32_t Size)
 {
 	if (!Buf)
 		return 0;
-	Buf->Data = OS_Zalloc(Size);
+	Buf->Data = zalloc(Size);
 	if (!Buf->Data)
 	{
 		Buf->MaxLen = 0;
@@ -260,7 +211,7 @@ void OS_DeInitBuffer(Buffer_Struct *Buf)
 {
 	if (Buf->Data)
 	{
-		OS_Free(Buf->Data);
+		free(Buf->Data);
 	}
 	Buf->Data = NULL;
 	Buf->MaxLen = 0;
@@ -274,9 +225,9 @@ int32_t OS_ReInitBuffer(Buffer_Struct *Buf, uint32_t Size)
 
 	if (Buf->Data)
 	{
-		OS_Free(Buf->Data);
+		free(Buf->Data);
 	}
-	Buf->Data = OS_Zalloc(Size);
+	Buf->Data = zalloc(Size);
 	if (!Buf->Data)
 	{
 		Buf->MaxLen = 0;
@@ -309,7 +260,7 @@ int32_t OS_ReSizeBuffer(Buffer_Struct *Buf, uint32_t Size)
 //	if (Old)
 //	{
 //		memcpy(New, Old, Buf->Pos);
-//		OS_Free(Old);
+//		free(Old);
 //	}
 	uint32_t Critical = OS_EnterCritical();
 	New = bgetr(Buf->Data, Size);
@@ -337,7 +288,7 @@ int32_t OS_BufferWrite(Buffer_Struct *Buf, void *Data, uint32_t Len)
 	}
 	if (!Buf->Data)
 	{
-		Buf->Data = OS_Zalloc(Len);
+		Buf->Data = zalloc(Len);
 		if (!Buf->Data)
 		{
 			return -ERROR_NO_MEMORY;
@@ -371,7 +322,7 @@ int32_t OS_BufferWriteLimit(Buffer_Struct *Buf, void *Data, uint32_t Len)
 	}
 	if (!Buf->Data)
 	{
-		Buf->Data = OS_Zalloc(Len);
+		Buf->Data = zalloc(Len);
 		if (!Buf->Data)
 		{
 			return -ERROR_NO_MEMORY;
@@ -846,56 +797,6 @@ __attribute__((weak)) void OS_ExitCritical(uint32_t Critical)
 	}
 }
 
-__attribute__((weak)) void *OS_Malloc(uint32_t Size)
-{
-	void *p;
-	uint32_t Critical = OS_EnterCritical();
-	p = bget(Size);
-	OS_ExitCritical(Critical);
-	return p;
-}
-
-__attribute__((weak)) void *OS_Zalloc(uint32_t Size)
-{
-	void *p;
-	uint32_t Critical = OS_EnterCritical();
-	p = bgetz(Size);
-	OS_ExitCritical(Critical);
-	return p;
-}
-
-__attribute__((weak)) void *OS_Calloc(uint32_t count, uint32_t eltsize)
-{
-	void *p;
-	uint32_t Critical = OS_EnterCritical();
-	p = bgetz(count * eltsize);
-	OS_ExitCritical(Critical);
-	return p;
-}
-
-__attribute__((weak)) void OS_Free(void *p)
-{
-	if (((uint32_t)p >= (uint32_t)(&__os_heap_start)) && ((uint32_t)p <= (uint32_t)(&__ram_end)))
-	{
-		uint32_t Critical = OS_EnterCritical();
-		brel(p);
-		OS_ExitCritical(Critical);
-	}
-
-}
-
-__attribute__((weak)) void *OS_Realloc(void *buf, uint32_t size)
-{
-	void *p;
-	uint32_t Critical = OS_EnterCritical();
-//	p = bget(size);
-//	memcpy(p, buf, size);
-//	brel(buf);
-	p = bgetr(buf, size);
-	OS_ExitCritical(Critical);
-	return p;
-}
-
 __attribute__((weak)) void OS_MemInfo(uint32_t *curalloc, uint32_t *totfree, uint32_t *maxfree)
 {
 	unsigned long  nget, nrel;
@@ -908,7 +809,7 @@ __attribute__((weak)) int32_t OS_InitBuffer(Buffer_Struct *Buf, uint32_t Size)
 {
 	if (!Buf)
 		return 0;
-	Buf->Data = OS_Zalloc(Size);
+	Buf->Data = zalloc(Size);
 	if (!Buf->Data)
 	{
 		Buf->MaxLen = 0;
@@ -924,7 +825,7 @@ __attribute__((weak)) void OS_DeInitBuffer(Buffer_Struct *Buf)
 {
 	if (Buf->Data)
 	{
-		OS_Free(Buf->Data);
+		free(Buf->Data);
 	}
 	Buf->Data = NULL;
 	Buf->MaxLen = 0;
@@ -938,9 +839,9 @@ __attribute__((weak)) int32_t OS_ReInitBuffer(Buffer_Struct *Buf, uint32_t Size)
 
 	if (Buf->Data)
 	{
-		OS_Free(Buf->Data);
+		free(Buf->Data);
 	}
-	Buf->Data = OS_Zalloc(Size);
+	Buf->Data = zalloc(Size);
 	if (!Buf->Data)
 	{
 		Buf->MaxLen = 0;
@@ -973,7 +874,7 @@ __attribute__((weak)) int32_t OS_ReSizeBuffer(Buffer_Struct *Buf, uint32_t Size)
 //	if (Old)
 //	{
 //		memcpy(New, Old, Buf->Pos);
-//		OS_Free(Old);
+//		free(Old);
 //	}
 	uint32_t Critical = OS_EnterCritical();
 	New = bgetr(Buf->Data, Size);
@@ -1001,7 +902,7 @@ __attribute__((weak)) int32_t OS_BufferWrite(Buffer_Struct *Buf, void *Data, uin
 	}
 	if (!Buf->Data)
 	{
-		Buf->Data = OS_Zalloc(Len);
+		Buf->Data = zalloc(Len);
 		if (!Buf->Data)
 		{
 			return -ERROR_NO_MEMORY;
@@ -1035,7 +936,7 @@ __attribute__((weak)) int32_t OS_BufferWriteLimit(Buffer_Struct *Buf, void *Data
 	}
 	if (!Buf->Data)
 	{
-		Buf->Data = OS_Zalloc(Len);
+		Buf->Data = zalloc(Len);
 		if (!Buf->Data)
 		{
 			return -ERROR_NO_MEMORY;
@@ -2223,3 +2124,185 @@ uint32_t unicode_to_utf8(void *in, uint32_t unicodelen, uint8_t *out, uint8_t is
 }
 
 #endif
+
+void *__wrap_malloc(size_t Size)
+{
+	void *p;
+	uint32_t Critical = OS_EnterCritical();
+	p = bget(Size);
+	OS_ExitCritical(Critical);
+	return p;
+}
+
+void *__wrap_zalloc(size_t Size)
+{
+	void *p;
+	uint32_t Critical = OS_EnterCritical();
+	p = bgetz(Size);
+	OS_ExitCritical(Critical);
+	return p;
+}
+
+void *__wrap_calloc(size_t count, size_t eltsize)
+{
+	void *p;
+	uint32_t Critical = OS_EnterCritical();
+	p = bgetz(count * eltsize);
+	OS_ExitCritical(Critical);
+	return p;
+}
+
+void __wrap_free(void *p)
+{
+	if (((uint32_t)p >= (uint32_t)(&__os_heap_start)) && ((uint32_t)p <= (uint32_t)(&__ram_end)))
+	{
+		uint32_t Critical = OS_EnterCritical();
+		brel(p);
+		OS_ExitCritical(Critical);
+	}
+
+}
+
+void *__wrap_realloc(void *buf, size_t size)
+{
+	void *p;
+	uint32_t Critical = OS_EnterCritical();
+//	p = bget(size);
+//	memcpy(p, buf, size);
+//	brel(buf);
+	p = bgetr(buf, size);
+	OS_ExitCritical(Critical);
+	return p;
+}
+
+void *__wrap__malloc_r(size_t Size)
+{
+	void *p;
+	uint32_t Critical = OS_EnterCritical();
+	p = bget(Size);
+	OS_ExitCritical(Critical);
+	return p;
+}
+
+void *__wrap__zalloc_r(size_t Size)
+{
+	void *p;
+	uint32_t Critical = OS_EnterCritical();
+	p = bgetz(Size);
+	OS_ExitCritical(Critical);
+	return p;
+}
+
+void *__wrap__calloc_r(size_t count, size_t eltsize)
+{
+	void *p;
+	uint32_t Critical = OS_EnterCritical();
+	p = bgetz(count * eltsize);
+	OS_ExitCritical(Critical);
+	return p;
+}
+
+void __wrap__free_r(void *p)
+{
+	if (((uint32_t)p >= (uint32_t)(&__os_heap_start)) && ((uint32_t)p <= (uint32_t)(&__ram_end)))
+	{
+		uint32_t Critical = OS_EnterCritical();
+		brel(p);
+		OS_ExitCritical(Critical);
+	}
+
+}
+
+void *__wrap__realloc_r(void *buf, size_t size)
+{
+	void *p;
+	uint32_t Critical = OS_EnterCritical();
+//	p = bget(size);
+//	memcpy(p, buf, size);
+//	brel(buf);
+	p = bgetr(buf, size);
+	OS_ExitCritical(Critical);
+	return p;
+}
+
+void __wrap___assert (const char *s1, int d1, const char *s2)
+{
+	__disable_irq();
+	DBG_Trace("assert %s,%d,%s", s1, d1, s2);
+	for( ;; );
+}
+
+void __wrap___assert_func (const char *s1, int d1, const char *s2, const char *s3)
+{
+	__disable_irq();
+	DBG_Trace("assert %s,%d,%s,%s", s1, d1, s2, s3);
+	for( ;; );
+}
+
+static struct tm prvTM;
+
+struct tm *__wrap_localtime (const time_t *_timer)
+{
+	Time_UserDataStruct Time;
+	Date_UserDataStruct Date;
+	uint64_t Sec;
+	if (_timer)
+	{
+		Sec = *_timer;
+		Tamp2UTC(Sec, &Date, &Time, 0);
+	}
+	else
+	{
+		RTC_GetDateTime(&Date, &Time);
+	}
+	prvTM.tm_year = Date.Year - 1900;
+	prvTM.tm_mon = Date.Mon - 1;
+	prvTM.tm_mday = Date.Day;
+	prvTM.tm_hour = Time.Hour;
+	prvTM.tm_min = Time.Min;
+	prvTM.tm_sec = Time.Sec;
+	return &prvTM;
+}
+
+struct tm *__wrap_gmtime (const time_t *_timer)
+{
+	Time_UserDataStruct Time;
+	Date_UserDataStruct Date;
+	uint64_t Sec;
+	if (_timer)
+	{
+		Sec = *_timer;
+		Tamp2UTC(Sec, &Date, &Time, 0);
+	}
+	else
+	{
+		RTC_GetDateTime(&Date, &Time);
+	}
+	prvTM.tm_year = Date.Year - 1900;
+	prvTM.tm_mon = Date.Mon - 1;
+	prvTM.tm_mday = Date.Day;
+	prvTM.tm_hour = Time.Hour;
+	prvTM.tm_min = Time.Min;
+	prvTM.tm_sec = Time.Sec;
+	return &prvTM;
+}
+
+clock_t	   __wrap_clock (void)
+{
+	return GetSysTickMS()/1000;
+}
+
+time_t	   __wrap_time (time_t *_timer)
+{
+	time_t t = RTC_GetUTC();
+	if (_timer)
+	{
+		*_timer = t;
+	}
+	return t;
+}
+
+void	_exit (int __status)
+{
+	return;
+}

bsp/common/src/core_task.c

@@ -89,7 +89,7 @@ HANDLE Task_Create(TaskFun_t EntryFunction, void *Param, uint32_t StackSize, uin
 		return NULL;
 	}
 	if (!Handle) return NULL;
-	Head = OS_Zalloc(sizeof(llist_head));
+	Head = zalloc(sizeof(llist_head));
 #ifdef __USE_SEMAPHORE__
 	Sem = OS_MutexCreate();
 	DelaySem = OS_MutexCreate();
@@ -122,7 +122,7 @@ void Task_SendEvent(HANDLE TaskHandle, uint32_t ID, uint32_t P1, uint32_t P2, ui
 #ifdef __USE_SEMAPHORE__
 	SemaphoreHandle_t sem = (SemaphoreHandle_t)vTaskGetPoint(TaskHandle, TASK_POINT_EVENT_SEM);
 #endif
-	Core_EventStruct *Event = OS_Zalloc(sizeof(Core_EventStruct));
+	Core_EventStruct *Event = zalloc(sizeof(Core_EventStruct));
 	Event->ID = ID;
 	Event->Param1 = P1;
 	Event->Param2 = P2;
@@ -207,7 +207,7 @@ GET_NEW_EVENT:
 		OutEvent->Param1 = Event->Param1;
 		OutEvent->Param2 = Event->Param2;
 		OutEvent->Param3 = Event->Param3;
-		OS_Free(Event);
+		free(Event);
 	}
 	else
 	{

xmake.lua

@@ -69,7 +69,35 @@ add_asflags("-mcpu=cortex-m4","-mfpu=fpv4-sp-d16","-mfloat-abi=hard","-mthumb","
 -- add_arflags("-mcpu=cortex-m4","-mfpu=fpv4-sp-d16","-mfloat-abi=hard","-mthumb","-c","--specs=nano.specs","-ffunction-sections","-fdata-sections","-fstack-usage","-Og","-DTRACE_LEVEL=4",{force = true})
 add_cxflags("-mcpu=cortex-m4","-Os","-mfpu=fpv4-sp-d16","-mfloat-abi=hard","-mthumb","-c","--specs=nano.specs","-ffunction-sections","-fdata-sections","-fstack-usage","-DTRACE_LEVEL=4",{force = true})
 
-add_ldflags("-mcpu=cortex-m4","-Os","-mfpu=fpv4-sp-d16","-mfloat-abi=hard","-mthumb","--specs=nano.specs","--specs=nosys.specs","-Wl,--gc-sections","-Wl,--check-sections","-Wl,--cref","-Wl,--no-whole-archive","-lc_nano","-Wl,--no-whole-archive",{force = true})
+add_ldflags("-mcpu=cortex-m4","-Os","-mfpu=fpv4-sp-d16","-mfloat-abi=hard","-mthumb","-static","--specs=nano.specs","-Wl,--gc-sections","-Wl,--check-sections","-Wl,--cref","-Wl,--no-whole-archive","-lc_nano","-Wl,--no-whole-archive",{force = true})
+
+
+add_ldflags(" -Wl,--wrap=malloc ",{force = true})
+add_ldflags(" -Wl,--wrap=free ",{force = true})
+add_ldflags(" -Wl,--wrap=zalloc ",{force = true})
+add_ldflags(" -Wl,--wrap=calloc ",{force = true})
+add_ldflags(" -Wl,--wrap=realloc ",{force = true})
+add_ldflags(" -Wl,--wrap=_malloc_r ",{force = true})
+add_ldflags(" -Wl,--wrap=_free_r ",{force = true})
+add_ldflags(" -Wl,--wrap=_zalloc_r ",{force = true})
+add_ldflags(" -Wl,--wrap=_calloc_r ",{force = true})
+add_ldflags(" -Wl,--wrap=_realloc_r ",{force = true})
+add_ldflags(" -Wl,--wrap=printf ",{force = true})
+add_ldflags(" -Wl,--wrap=sprintf ",{force = true})
+add_ldflags(" -Wl,--wrap=snprintf ",{force = true})
+add_ldflags(" -Wl,--wrap=vprintf ",{force = true})
+add_ldflags(" -Wl,--wrap=vsnprintf ",{force = true})
+add_ldflags(" -Wl,--wrap=fprintf ",{force = true})
+add_ldflags(" -Wl,--wrap=__assert ",{force = true})
+add_ldflags(" -Wl,--wrap=__assert_func ",{force = true})
+add_ldflags(" -Wl,--wrap=clock ",{force = true})
+add_ldflags(" -Wl,--wrap=localtime ",{force = true})
+add_ldflags(" -Wl,--wrap=gmtime ",{force = true})
+add_ldflags(" -Wl,--wrap=time ",{force = true})
+add_ldflags(" -Wl,--wrap=putc ",{force = true})
+add_ldflags(" -Wl,--wrap=putchar ",{force = true})
+add_ldflags(" -Wl,--wrap=puts ",{force = true})
+
 
 set_dependir("$(buildir)/.deps")
 set_objectdir("$(buildir)/.objs")
@@ -235,7 +263,12 @@ end
     add_files("Third_Party/jpeg_encode/*.c",{public = true})
     add_includedirs("Third_Party/jpeg_encode",{public = true})
 
-    
+    add_files("Third_Party/heap/*.c",{public = true})
+    add_includedirs("Third_Party/heap",{public = true})    
+
+    add_files("Third_Party/vsprintf/*.c",{public = true})
+    add_includedirs("Third_Party/vsprintf",{public = true})    
+
     --add_files("bsp/common/*.c",{public = true})
 	add_files("bsp/common/src/*.c",{public = true})
     add_includedirs("bsp/cmsis/include",{public = true})
@@ -272,6 +305,8 @@ if with_luatos then
     add_files("application/src/*.c")
 
     add_files(luatos.."lua/src/*.c")
+    remove_files(luatos.."lua/src/printf.c")
+    remove_files(luatos.."lua/src/bget.c")
     add_files(luatos.."luat/modules/*.c")
     remove_files(luatos.."luat/modules/luat_lib_http.c")
     add_files(luatos.."luat/vfs/*.c")
@@ -395,6 +430,10 @@ if with_luatos then
 
     -- crypto
     add_files(luatos.."components/crypto/**.c")
+
+    -- protobuf
+    add_includedirs(luatos.."components/serialization/protobuf")
+    add_files(luatos.."components/serialization/protobuf/*.c")
 else
 
     add_files("Third_Party/vsprintf/*.c",{public = true})