Reconfigurable robotic glove for human-in-the-loop rehabilitation protocols

How can a robotic glove enhance rehabilitation by integrating with the human neuromuscular system? While soft robotics and EMG-based control have advanced assistive technology, many systems lack adaptability and multimodal capabilities. This thesis presents a reconfigurable robotic glove for EMG-driven rehabilitation, designed to interface with various pneumatic devices while adapting to individual needs. The system employs three control modalities—assistance, training, and reward—to create a personalized rehabilitation process. An EMG-based strategy dynamically adjusts actuation, ensuring intuitive interaction where assistance decreases as the user’s strength improves. A key innovation is a universal control architecture applicable to any pneumatic rehabilitation device. Unlike conventional solutions tied to specific actuators, this framework extends beyond hand rehabilitation, as demonstrated with the TangiBall, a pneumatic device for grip training. This modular approach enables broader applications, including limb therapy. The glove is based on Digits, a reconfigurable pneumatic haptic interface. Iterative refinements optimized its design for comfort and biomechanical accuracy. The final model combines 3D-printed rigid components with soft pneumatic actuators, providing controlled motion for a natural and adaptive user experience. Experimental validation assessed its impact on motor recovery and adaptability. Results confirm that the control strategy effectively regulates user effort, bridging neuromuscular intent and robotic actuation. Tests with multiple users highlight the system’s robustness across different profiles, reinforcing its suitability for personalized rehabilitation. This research advances robotic rehabilitation by introducing a scalable, multimodal, and adaptive system. The integration of EMG-driven control, pneumatic actuation, and modular design sets a new benchmark for soft wearable robotics.