The IoT based PPG Signal Classification System for Acute Audio-Visual Stimulus Induced Stress

Abstract

Mental stress causes a great impact on our autonomic nervous system. Pulse rate variability (PRV) is a method which measures the changes in the autonomic nervous system of an individual. This study aims to acquire PPG signals in real time from a single spot Pulse sensor and then PRV analysis is performed on Pulse signal to determine perceived stress from the subject caused due to nerve-wracking audio-visual stimulus. PPG signal is then transferred wirelessly over an Android app. Also this work incorporates several Machine Learning models to organize the stress level of the subjects as average stress or high stress. Non-linear model gives best average classification precision, sensitivity and specificity of 90%, 100% and 82% respectively. With the advancement of portable PPG monitoring device acts as a substitute to Heart rate variability (HRV) even during the moving conditions. Also PPG signal is compared with ECG signal and a close precision is obtained with average percentage error of 8% for BPM and 3% for RR interval. PPG sensors offer more comfortness to the users which can be positioned on fingertip and wrist. By means of the improvement of android app provides feasibility to monitor stress by providing an alert to the mobile users whenever the stress exceeds the normal limits.