Embedded heart rate analysis based on sound sensing

Abstract

The exact location of heartbeats in the phonocardiogram signal is crucial for correct segmentation and classification of the S1 and S2 heart sounds. This task becomes difficult due to the inclusion of noise in the acquisition process and due to the fact that the implementation concerns the embedded system. In this article, we present a low-cost system of location and classification of heart sounds in S1 and S2 based on a single-chip microcontroller and MEMS microphone. The experimental data analysis methods in time and frequency domains are also presented. The heartbeat segmentation process includes autocorrelation to predict the time of the heartbeat cycle. The time-frequency characteristics are extracted with a Fast Fourier Transform to analyze diastole and systole heart cycles.