The Segmented Dynamic Time Warping Algorithm for Beat-to-Beat Heart Rate Estimation based on Ballistocardiogram Signals

Abstract

In order to improve the accuracy and speed of beat-to-beat heart rate estimation, a Segmented Dynamic Time Warping (SDTW) algorithm based on Ballistocardiogram (BCG) signal was presented. The beat-to-beat heart rate obtained through this algorithm was evaluated using 20 healthy subjects with synchronized lead I ECG as standard. The mean bias between JJ and RR interval was 0.2 ms and the confidence interval of 95% was ± 19 ms. It indicates that the obtained beat-to-beat intervals are in better agreement with that of ECG. The mean relative error and matching time for heart rate estimation with the algorithm were 1.37% and 0.77 s, respectively. The results were superior to that of the traditional template matching algorithms. It establishes the foundation for heart disease monitoring based on BCG signal.