Accurate Heart Rate Monitoring Method During Physical Exercise From Photoplethysmography Signal

Abstract

Photoplethysmography (PPG) is mainly used to find skin blood flow. Volumetric change of blood flow indicates cardiac rhythm, due to that there is a possibility to monitor the heart rate (HR) from PPG signals. PPG sensors can be implemented in wrist band or smart watch. HR monitoring during physical exercise can give helpful information regarding physical fitness tracking features like calorie measurement, sleep tracking, counting steps and help the trainer to achieve their training goals. Body movements during physical exercise produces noise content in PPG signal. This noise content is generally called as moving artifact. Monitoring accurate heart rate from PPG signal is a challenging task. Nowadays there are many methods are proposed to improve the HR monitoring from PPG signals. This paper is envisaged to develop a method to detect accurate HR without any complexity. The proposed work consists of two steps: 1) extraction of PPG signal from the observed PPG by sparse method and2) HR estimation. The main advantage of sparse method is the proper reconstruction of signals from small number of highly noisy elementary signals. This method eliminates the need of accelerometer reading and uses only two PPG signals, which results less complexity. Estimation of HR through the phase information, which provided by the polar representation of PPG signal. The proposed method results in an average absolute error of 1.00 BPM and average relative error of 0.85 BPM using 2 PPG signals compared to existing methods.