Bat-LMS based Adaptive Filtering and Stock-well Transform Based Characteristics Point Extraction for Impedance Cardiography

The automatic extraction of characteristic points in an Impedance Cardiography (ICG) signal is the critical step in any programmed ICG signal investigation and if the ICG signal is taken from the humans additionally contains some noise like signals, for example, motion artifacts. Extraction of ICG signals characteristic points and investigating the cardiac disorders of the human is a demanding task, since one single device is utilized and it gets a blend of different heart beats. In order to resolve such problems we propose a Bat-LMS based Adaptive filtering method with Stockwell Transform in this paper. The adaptive filter coefficients calculated by combining Bat algorithm with LMS algorithm. In this setup, error signal obtained from the filter is given to Bat algorithm which decides the proper step size with less error. Then the optimized step size value is given to LMS gives where it updates the coefficients of filter simultaneously. Finally, an automatic extraction of the characteristic points by utilizing S-Transform to detect start point (B), peak point (Z) and end point (X) of left ventricular ejection time accurately. The experimental results of ICG signal analysis shows the importance of our proposed adaptive filtering method.