Multi-sensor fusion for wearable heart rate monitoring system

Abstract:

To improve the accuracy of heart rate (HR) in daily behaviors, multi-sensor fusion method was used in this paper to fuse ECG and pulse wave (PW)whichis closely related to biological electrophysiology and biomechanics, respectively. And a wearable heart rate monitoring system with high reliability based on Android platform was achieved. The proposed system and ST-1212 ECG workstation were used for 18 cases simultaneousexperiment of different motion intensity in daily behaviors. Signal quality indices (SQI) that reflect the level of signal quality were calculated by analyzing the signal characteristics in time domain, and then Kalman-Filter (KF) was adaptively regulated to make the optimal estimation of the HR derivedfrom the dual-channel signal according to SQI, and finally KF residuals were used to adjust the weights to get the fused HR.The results indicate that the fused HR can improve the accuracy more than 46% than those derived from ECG or PW directly. The system can effectively reduce the artifact on HR estimationby using multi-sensor fusion method, thus it can be used for continuous monitoring of HR with low physiological and mental burden for a relatively long time.