A comparison of three heart rate detection algorithms over ballistocardiogram signals

Heart rate (HR) detection from ballistocardiogram (BCG) signals is challenging because the signal morphology can vary between and within-subjects. Also, it differs one sensor to another. Hence, essential evaluate HR algorithms across several datasets under different experimental setups. In this paper, we studied potential of three four independent BCG datasets. The are as follows: multiresolution analysis maximal overlap discrete wavelet transform (MODWT-MRA), continuous (CWT), template matching (TM). were obtained using a microbend fiber optic sensor, Bragg grating electromechanical films, load cells, respectively. gathered from: a) 10 patients during polysomnography study, b) 50 subjects in sitting position, c) sleeping d) 40 position. Overall, CWT with derivative Gaussian provided superior results compared MODWT-MRA, (frequency B-spline), (Shannon). That said, was constructed DataSet1. Then, used for other TM method achieved satisfactory DataSet2 DataSet3, but did not detect two DataSet4. proposed methods implemented on Raspberry Pi. As result, average time required analyze 30-second less than second all methods. Yet, MODWT-MRA had highest performance an 0.04 seconds.