A modified approach of peak extraction from BCG for heart rate estimation

Abstract

We propose a novel method for detecting a single heartbeat interval from ballistocardiogram (BCG) from healthy subjects. An inconspicuous tilt sensor is embedded in the mattress to record a robust estimate of the local beat interval. The sensor could measure body movements caused by cardiac activity. Compared to many existing methods, this new method provides real-time and accurate heart rate estimation for all normal data and most both artifacts and anamorphic data on a beat-by-beat basis. The consistency of the proposed method with the ECG reference has been evaluated. In this data set containing approximately 15,000 heart beats, records of 10 subjects achieved an average beat interval error of 0.88% with a coverage rate of 96.94%.