Adaptive weight compensation in Assistive Upper-Limb Exoskeletons: an EMG Analysis

Abstract

This study explores the integration of adaptive weight compensation within an upper-limb exoskeleton by analyzing muscle activity from electromyography (EMG) signals. We enabled the exoskeleton to translate user-applied forces into precise joint movements by implementing an admittance controller. Through our experimental protocol, we measured EMG signals from the biceps brachii muscle, focusing on changes in two time-domain features -RMS and variance-under varying conditions. Then, we analyzed these changes to evaluate the impact on the muscle when users were assisted in carrying a one-kilogram object, compared to when they were not assisted. Our initial findings indicate that adaptive weight compensation significantly reduces muscle activity and enhances stability during lifting tasks, suggesting potential for real-time EMG-based activation of weight compensation. This research highlights the importance of combining force and EMG data to develop adaptive control laws for assistive robotics, paving the way for personalized Human-Robot Interaction (HRI).