Power line noise removal from ECG signal using notch, band stop and adaptive filters

Abstract

Heart rhythms can be recorded and electrically displayed using Electrocardiogram (ECG) signal with non-stationary properties. However, such recordings can be corrupted by different types of proximity noise including power line interference (PLI). The corruption can cause variation in amplitude and/or duration of signal leading to diagnostic errors. Traditional notch filters are ineffective against such noise due to non-static nature of PLI frequency. In this paper, we study ECG signal de-noising technique using the traditional notch filter, band stop filter, Least-Mean-Square (LMS) filter. Results are compared both in frequency and time domain. The program is written using Signal Processing Toolbox (MATLAB). Our results show that LMS filter provides superior results in comparison to other methods for de-noising of ECG signals. Keywords: Electrocardiography, Adaptive filters, Notch filters, IIR filters, Interference, Band-pass filters, Finite impulse response filters