Hydrogel-Based Self-powered Sensing for Muscle Activity Monitoring and Safe Rehabilitation Robotics

Abstract Upper-limb motor dysfunction is a common sequela following neurological and orthopedic injuries and can severely impair functional recovery. However, conventional rehabilitation approaches remain limited by subjective assessments and the lack of muscle-specific physiological feedback. In this study, a self-powered adaptive coupled sensing system based on a flexible triboelectric nanogenerator (TENG) is developed. The TENG is constructed using a polyvinyl alcohol/sodium alginate/polyacrylamide double-network organohydrogel. This hydrogel serves as both a highly deformable electrode and an efficient tribopositive layer. It exhibits a tensile strength of 400 kPa, an elongation at break of 325%, and excellent ionic conductivity. The fabricated TENG sensors, including PP-TENG and PSPE-TENG, maintain stable electrical output and rapid response performance even at a low operating frequency of 1 Hz. The response and recovery times are approximately 84 and 74 ms, respectively. Owing to their excellent conformability, the sensors can be attached directly to the skin and dynamically monitor subtle muscle deformations during movement. Furthermore, the PP-TENG was integrated with a Universal Robots UR7e robotic arm to establish a closed-loop control framework. The system continuously evaluates muscle contraction stability through real-time analysis of the variance of TENG signals. Once abnormal fluctuations in muscle activity are detected, the system automatically suspends robotic assistance and issues a warning, enabling rapid safety monitoring during rehabilitation training. Proof-of-concept demonstrations in biceps-curl exercises and ankle rehabilitation showed that the system can adaptively regulate training intensity based on direct neuromuscular feedback. This work provides a promising strategy for hydrogel-based self-powered sensing and offers a new pathway toward safe and personalized closed-loop rehabilitation systems.