Cost-Effective EMG Signal Acquisition for Rehabilitation Robotics using Single-Channel Sensors and Machine Learning
Nakul Sanwlot, K. Vishnu Raj, Ganesh Udupa, R Anand
Abstract
Electromyography study focuses on the classification of electromyography (EMG) signals using machine learning (ML) techniques. The classification of EMG signals with ML techniques improves the response and accuracy of myo-electric prosthetic hands and rehabilitation systems. One of the key techniques to get reliable surface EMG signals is to employ multiple sensors. Integrating multiple sensors in myoelectric hands can inflate the cost of prosthetic hands. To study the effectiveness of implementing a single sensor combined with machine learning algorithms to accurately classify the EMG signals, an experiment was conducted and reported in the paper. EMG signals were collected from five volunteers to create a comprehensive dataset. Various machine learning algorithms were then applied to classify the EMG data effectively. The proposed approach achieved an accuracy of over 80%, highlighting the potential of using single-channel EMG sensors combined with machine learning for accurate signal classification. This approach could be beneficial for various applications, including prosthetic control and rehabilitation devices.