WIRELESS COMMUNICATION BASED HEART RATE DETECTION USING WEARABLE DEVICE

Estimation of heart rate is a challenging task from PPG (photoplethysmographic) signal, as motion artifacts gets included during physical activities. The focus of this research work is to automatically estimate the heart rate of the person based on digital signal processing, where this approach is able to classify, process and analyze vital physiological signals for the application of health monitoring in long term. This work presents a novel approach, exploiting Adaptive Wiener filter for attenuating and removal of motion artifacts with a phase vocoder for refining the estimate of Heart Rate (HR). Two PPG signals and three accelerometer signals are obtained from the wearable device in the form of wrist band, are first preprocessed before applying to the adaptive wiener filter. Physionet database that is available publically with 12 PPG recordings is considered for analyzing the performance of the novel estimation of HR system. Ultimately, heart rate estimation is done based on subsequent detection algorithm. MATLAB implementation of the proposed method is able to achieve an average absolute error of 1.08 BPM (Beats per Minute). Correlation coefficient between estimated heart rate and true heart rate is obtained as 0.997. Fine tuning of spectral peak is achieved in this work which is accurately tracked for estimation of heart rate