EMG-Based Control of a Mobile Robot for Upper Limb Rehabilitation

This work presents the design and implementation of a proposal for a physical rehabilitation system for upper limbs, based on the use of surface EMG sensors that record the activity of four muscles: middle deltoid, pronator teres, biceps brachii, and brachioradialis. Based on real-time processing of these signals, each muscle activation generates a command that allows the movement of a TurtleBot3, through ROS, within a controlled environment. The experiment was conducted with 20 participants distributed in an experimental group and a control group, evaluating the trajectory, distance, and time metrics generated by the robot in three round-trip trials on a predefined trajectory. The results show a decrease in trajectory dispersion and a reduction in execution time, indicating a learning process in muscle use between repetitions. The mobile robot control through EMG signals is concluded to be viable as a therapeutic tool in physical rehabilitation processes.