Determination of Absolute Heart Beat from Photoplethysmographic Signals in the Presence of Motion Artifacts

In Wireless Body Area Networks (WBANs), accurate monitoring of heart rate (HR) using Photoplethysmography (PPG) signals is always a difficult task, especially when the subjects are under radical exercises. This is due to the signals corrupted by severely strong Motion Artifacts (MA) caused by the subject's body movements. In this work, a novel approach has been proposed consisting of signal decomposition for denoising using principal component analysis (PCA), spare signal reconstruction (SSR), peak detection and tracking and support vector machine (SVM) classifier for accurate estimation of HR, based on the wrist type PPG signals. With this approach, we are able to achieve high accuracy and also, it is strong enough to remove MA. Experiments were conducted on 12 subjects and their datasets are obtained from 2015 IEEE Signal Processing CUP, running on a threadmill with varying speeds ranging from 0 to a maximum speed of 15 km/hour. From the results, it is observed that the average absolute error of heart rate estimation is 1.66 beats per minute (BPM).