Analysis of Parameters for Smoothing Electrocardiographic Signals

The article is devoted to the consideration of the features of smoothing of ECG signal against the background of electromyographic distortions of various magnitude. The main goal of the research is comparative analysis of various options for the implementation of smoothing of an ECG signal contaminated by myographic interference in order to determine the optimal approach in terms of minimizing biosignal distortions and measurement errors of its amplitude-time characteristics. To obtain quantitative characteristics of the effectiveness of various methods for smoothing of the ECG signal, an approach was used based on simulation models of the ECG signal and distortions. A criterion for choosing the optimal parameters for ECG signal smoothing based on minimizing the errors in determining the durations of RR-intervals and distortions of the ECG signal was proposed. Various options for smoothing filters are considered: low-pass filter, multiscale wavelet transform, Savitzky–Golay filter, moving average filter. The optimal parameters for each type of filter are determined in terms of minimizing the distortion of the ECG signal and the measurement error of the durations of RR-intervals. The dependences of the change in the measurement error of the durations of RR-intervals on the signal-to-noise ratio, the dependences of the change in the signal distortion coefficient on the signal-to-noise ratio, and the plots of processing the noisy fragment of ECG signal by various types of filters are presented. Studies have shown that multiscale wavelet transforms of ECG signal with myographic interference is the optimal method for processing an ECG signal, providing minimal measurement errors of RR-intervals with minimal distortion of the ECG signal.