A Novel Low-Pressure Robotic Glove Based on CT-Optimized Finger Joint Kinematic Model for Long-Term Rehabilitation of Stroke Patients

Wearing robotic gloves has become increasingly crucial for hand rehabilitation in stroke patients. However, traditional robotic gloves can exert additional pressure on the hand, such as prolonged use leading to poor blood circulation and muscle stiffness. To address these concerns, this work analyzes the finger kinematic model based on computerized tomography (CT) images of human hands, and designs a low-pressure robotic glove that conforms to finger kinematic characteristics. Firstly, physiological data on finger joint flexion and extension were collected through CT scans. The equivalent rotation centers of finger joints were obtained using the SURF and RANSAC algorithms. Furthermore, the trajectory of finger joint end and the correlation equation of finger joint motion were fitted, and a comprehensive finger kinematic model was established. Based on this finger kinematic model, a novel under-actuated exoskeleton mechanism was designed using a human-machine integration approach. The novel robotic glove fully aligns with the equivalent rotation centers and natural motion trajectories of the fingers, exerting minimal and evenly distributed dynamic pressure on the fingers, with a theoretical static pressure value of zero. Experiments involving gripping everyday objects demonstrated that the novel robotic glove significantly reduces the overall pressure on the fingers during grasping compared to the pneumatic glove and the traditional exoskeleton robotic glove. It is suitable for long-term use by stroke patients for rehabilitation training.