Efficient ECG Signal Conditioning Techniques using Variable Step Size Least Mean Fourth Algorithms

Abstract—obtaining an artifact free signal is an important task in making a successful diagnosis using Electro Cardiograph (ECG) signal. Several techniques were proposed in the literature with varying degree of accuracy. In this paper some efficient signal conditioning techniques to remove the artifact from ECG signals are presented. The proposed techniques are derived from basic higher order technique known as of Least Mean Fourth (LMF) algorithm.All the techniques are evaluated using MIT-BIH arrhythmia database. The SNR performance of the techniques is calculated and is compared with Normalized Least Mean square (NLMS) algorithm. From the SNR measurements obtained variable XENLMF was found to be exhibiting the superior performance over the NLMS and the other techniques and on an average the SNR values of variable XE-NLMF in case of PLI, BW,MA and EM artifacts are 10.7800dB, 8.5950dB, 9.0703dB, 8.3210dB which are better than their counterparts. The convergence characteristics of all these techniques measured have further shown the suitability of the Variable XENLMF technique over the other in using at real time situations Keyword Electro Cardio Graph, Artifact, Least Mean Fourth Algorithm, Noise Cancellation.