A novel fixed-point leaky sign regressor algorithm based adaptive noise canceller for PLI cancellation in ECG signals

Abstract

In this paper, a novel fixed-point Leaky Sign Regressor Algorithm (LSRA) based adaptive noise canceller has been employed for the cancellation of 60 Hz Power Line Interference (PLI) from the ElectroCardioGram (ECG) signal. A sufficient condition for the convergence in the mean of the LSRA algorithm is also derived. The fixed-point LSRA-based adaptive noise canceller employed in this work is fully quantized using an in-house quantize function. The most effective number of quantization bits required for the various parameters are found to be 6-bits and are determined through rigorous simulations. The filtered ECG signal free from 60 Hz PLI is successfully recovered using a novel 6-bit fixed-point LSRA-based adaptive noise canceller.

A New Fixed Point Noise Cancellation Method for Suppressing Power Line Interference in Electrocardiogram Signals

“…Note that the expression for the step size upper bound of the LSRLMS algorithm in ( 9) is exactly the same as that derived in [12].…”

“…The proposed LSRLMMN algorithm is used in the fixed point noise cancellation technique described in [12], [13]. The LSRLMMN algorithm is a mixture of the LSRLMS and Leaky Sign Regressor Least Mean Fourth (LSRLMF) algorithms as long as the mixing variable ∆ is in the range of 0 < ∆ < 1.…”

“…Additionally, the step size upper bounds of the LLMMN algorithm was derived [11]. A new fixed point Leaky Sign Regressor Least Mean Square (LSRLMS) powered noise cancellation technique was proposed for eliminating the 60-Hz PLI noise embedded in the ECG signal [12]. Additionally, the step size upper bound of the LSRLMS algorithm was derived [12].…”

“…A new fixed point Leaky Sign Regressor Least Mean Square (LSRLMS) powered noise cancellation technique was proposed for eliminating the 60-Hz PLI noise embedded in the ECG signal [12]. Additionally, the step size upper bound of the LSRLMS algorithm was derived [12]. In this article, a new fixed point Leaky Sign Regressor Least Mean Mixed Norm (LSRLMMN) powered noise cancellation technique is proposed for the removal of 60-Hz PLI noise present in the ECG signal.…”