Stress Detection based on ECG, GSR Signals, HR, and Behavioral Context

Abstract

Stress has become a significant global issue in recent years. People of all ages experience stress level problems at various points in their lives. As technology develops at an ever-increasing rate, everyone falls behind the competition, which significantly affects human health by causing mental stress. As a result, early stress detection is necessary. So, we proposed a methodology to efficiently detect whether or not a person is stressed and provide a stress level. This work concentrates on finding a person's stress by using physiological signals such as Galvanic Skin Response (GSR), Electrocardiogram (ECG) levels, Heart Rate (HR), and behavioral actions that a user is usually doing during the day. Each physiological signal's data is collected from different resources and integrated, and analyzed, and a dataset with a healthy and stressed population is generated. This research focuses on applying deep learning algorithms to enhance the performance of the stress detection system. We have designed a webpage for stress detection which is user-friendly. People can perform the prediction analysis on the Stress Detection System using live HR data and GSR from the heart rate and Grove-GSR sensors connected to the Arduino. By using this application user can obtain information about their stress levels, which allows them to make the appropriate treatment decisions based on the results.