Artificial Intelligence in Control of an Actuated Glove for Hand Rehabilitation

Hand rehabilitation following neurological injuries requires intensive, repetitive, and task-specific training to restore fine motor function. Actuated gloves – wearable robotic devices that assist finger and hand movements – offer a compact and patient-friendly alternative to traditional robotic arms. This paper presents an AI-driven control framework for an actuated rehabilitation glove, focusing on adaptive assistance, intention recognition, and safe human–robot interaction. The system integrates multimodal sensing, including electromyography (EMG), flex sensors, and force feedback, with machine learning and reinforcement learning algorithms for personalized therapy. A dynamic model of the glove–hand system is introduced, and control strategies combining impedance control and AI-based adaptation are analyzed. The results indicate that AI-enhanced glove systems significantly improve motor recovery, user engagement, and control precision.