EMG-based evaluation of relationship between muscle activity, joint kinematics, and joint torque in an upper limb exoskeleton supporting range of motion exercises

Shoulder mobility impairments present substantial challenges in rehabilitation, where limited therapist availability and inconsistent exercise compliance frequently hinder recovery outcomes. Robotic exoskeletons offer a promising solution by delivering precise, repeatable mechanical assistance during unassisted therapy sessions. This paper presents the experimental validation of a two-degree-of-freedom active upper-limb exoskeleton prototype designed for sagittal-plane upper limb rehabilitation exercises. The system implements two operational modes: a constant joint torque mode and a gravity compensation mode. Muscle activity during different scenarios was assessed across ten healthy male participants using surface electromyography (sEMG) recorded from the anterior deltoid, posterior deltoid, biceps, and triceps during three movement tasks at six discrete torque levels. Processed sEMG signals were standardized via z-score normalization and characterized through linear regression against joint angle. Results demonstrated angle-dependent activation scaling in primary agonist muscles, which translated monotonically with opposing external torque. Additionally, gravity compensation systematically attenuated both agonist and antagonist muscle activity during anti-gravity trajectories. Substantial inter-subject variability, particularly during shoulder rotations, suggests that individualized physiological recruitment strategies exert a dominant influence on sEMG amplitude. These findings establish a predictive behavioral framework for designing targeted rehabilitation interventions with adjustable mechanical loading profiles without using complex biosensors.