A tolerant algorithm for cardiac pulses characterization in ballistocardiography signals in a non-invasive system

A variety of signal processing methods have been developed for detecting the heartbeat from signals obtained by non-invasive devices. In this paper, an algorithm developed for the accurate detection of heartbeats from a ballistocardiography signal is presented. The signal is obtained using a regular chair adapted with Electro-Mechanical Film (EMFi) sensors and an acquisition system, so a ballistocardiogram is generated. The presence of important disturbances in this type of signal is unavoidable due to several natural causes; therefore, its processing involves significant difficulties to be tackled in order to obtain reliable lectures of the heartbeat. The algorithm presented here performs consistently well, showing a very low error rate, even if the signal contains significant noise.