Heart rate estimation using an EMG system integrated with a motion capture system

Abstract

In electromyography (EMG) data, the electrocardiography (ECG) signal frequently is also present. In many papers this signal is treated as a distortion and the focus is on its removal. However, the ECG signal can be useful. For example, if an athlete is subject to a stress test exercise in a motion capture laboratory, his/her current heart rate synchronized with recorded motion may provide valuable additional information. In this paper we propose a method for computing an athlete's heart rate using the EMG signal. The method is not computationally expensive. Detected heart beats can be saved in a standard C3D file as an event. This makes further examination of the correlations between an athlete's movement and his/her heart rate much easier. A number of trials were conducted to verify the method. Data was recorded in a resting position (sitting on a chair), during walking and during exercise (lifting on one's hands). The goals of the trials were the verification of the method's correctness (resting test) and the checking of how much movement affects the accuracy of the measurements (witch special focus on other muscles near the heart). Measurements were made in parallel with the Vershold pulse meter and the 16-channel wireless EMG system from Myon, controlled by Vicon Nexus software. Unlike standard electrocardiographs, where multiple electrodes are used, only one voltage between two points was measured (due to the properties of the wireless EMG system used). For this reason, different electrode fixing variants were examined. In this way it could be determined how position affects the accuracy of measurement. The results show the advantage of our method for extracting ECG from EMG data over the pulse meter, which requires accurate positioning on the body. If it is not mounted properly the accuracy of the results obtained with it will drop significantly.