Recursive Root Mean Square estimator

Real-time digital systems often require the calculation of a root-mean, such as a root-mean-square (RMS) level or average magnitude of a complex signal. While averaging can be efficiently implemented by most microprocessors, obtaining an efficient and rapidly converging algorithm is surely a time-consuming task due to the adjustments and optimizations that need to be performed. On the other hand, calculating the square root can be challenging, especially with low-cost hardware. If the processor does not have a built-in fast square root function, it must be implemented in software. While this can provide accurate results, it may not be as efficient.

qlibs::rms provides a super-efficient method to compute the Recursive Root Mean Square (RMS). The implementation uses 3-stage filtering and Newton's method to obtain the square root. The filter parameters are perfectly tuned and optimized to converge in around 16 cycles, operating at a sampling rate of 1mS. This means that you must configure your embedded system to meet those specifications, either by configuring a timer interrupt or by delegating responsibility to a real-time task.

Example : By using a timer interrupt

rms_compute.c

#include "bsp.h"
#include "hal_timer.h"
#include <qlibs.h>
 
using namespace qlibs;
 
static rms rms_instance;
static volatile real_t rms_value = 0.0f;
static real_t window[ 8 ] = { 0.0f };
 
void TimerA0_Int_Handler( void ) //interrupt tick at 1mS
{
    rms_value = rms_instance.update( get_sample() );
    HAL_TimerA0_ClearStatusFlag();
}
 
void rms_setup( void )
{
    rms_instance.setup( window );
    HAL_TimerA0_Init();
    HAL_TimerA0_SetPeriod_mS( 1 );
    HAL_TimerA0_EnableInterrupts();
    HAL_TimerA0_Enable();
}
 
float rms_getvalue( void )
{
    return rms_value;
}

rms_compute.h

#ifndef RMS_COMPUTE_H
#define RMS_COMPUTE_H
 
#ifdef __cplusplus
extern "C" {
#endif
 
    void rms_setup( void );
    float rms_getvalue( void );
 
#ifdef __cplusplus
}
#endif
 
#endif