Heart Rate Prediction of Running Exercise Based on Neural Network Utilizing Predefined Intensity and Route Information

Heart rate is closely linked to exercise intensity. Heart rate prediction is a problem which can increase the safety and the effectiveness during exercise. For outdoor running, exercise intensity is affected by both running speed and route contour. This paper presents a heart rate prediction system that uses planned speed and route information. The system takes as inputs the current heart rate, planned speed, and route to forecast future heart rates, leveraging data collected from a smartwatch. It is built on a simple three-layer neural network, achieving high accuracy with a mean squared error (MSE) as low as 0.0013. An evaluation of the prediction time interval indicates that shorter intervals improve accuracy. Additionally, this paper proposes an exercise recommendation system that considers planned intensity, including running speed and route information.