Low-Cost Heart Rate Sensor and Mental Stress Detection Using Machine Learning

One of our major organs heart does the pumping process of oxygen-containing blood and its distribution to the body's arteries every minute. Heart rate or pulse indicates the cardiovascular fitness of a human body. The health condition is predicted by measuring the heartbeat rate, which changes with age, physical and mental conditions. The most familiar way of measuring the heart rate or rhythm is by sensing the pulse per minute by various devices. This paper implements a low-cost heart rate monitoring system using sensors and IoT devices. First, the sensor will be placed on the finger, and subsequently, the color variation will be seen. The sensor picks the color variation, and it measures the interval of color variation. An Arduino microcontroller is used to process the signal. These devices use light to track the blood. Next, the measured heart rate data from the Arduino is stored in CSV files. The Geneva affective picture database has been used to record the heart rate and classify it into three classes of positive, negative, and neutral emotions. Finally, a machine learning algorithm, support vector machine, has been implemented to predict the mental stress condition from the obtained heart rate. Experimental results demonstrate that the support vector machine with the polynomial kernel exhibits the best accuracy.