Unobtrusive Inter-beat Interval Estimation from Multichannel Ballistocardiogram Signal Using Kalman Filter

Unobtrusively detecting inter-beat interval (IBI) from ballistocardiogram (BCG) is useful for monitoring cardiac activity at home, especially for calculating heart rate variability (HRV), the critical indicator to evaluate heart health. Compared to single-sensor system in most studies, this research used a bed-embedded 9 by 2 array sensors system to improve measurement coverage and precision of IBI estimation. Based on this system, we proposed a mode-switch based algorithm to solve the problem on array sensor signal selection and multichannel data fusion using linear regression model and Kalman filter. In addition, a peak detection algorithm was designed to estimate IBI from each channel signal. The algorithm was validated by approximately 48 hours BCG recordings captured from 24 subjects with different sleeping positions. A mean absolute error of 31ms at 83% average coverage was obtained by the proposed method, which has proven to be a promising candidate for IBI estimation from BCG signal on multichannel array sensors system.