Assist-As-Needed Exoskeleton for Hand Joint Rehabilitation Based on Muscle Effort Detection

Area of Science:

• Rehabilitation Robotics
• Neurorehabilitation Engineering
• Biomedical Signal Processing

Background:

• Robotic-assisted systems are increasingly used in post-stroke rehabilitation for high-intensity, controlled therapy.
• Existing systems require precise motion control but often lack dynamic adaptation to patient effort.

Purpose of the Study:

• To develop a novel closed-loop robotic exoskeleton architecture for active rehabilitation support.
• To integrate electromyographic (EMG) signal measurement for real-time adjustment of exoskeleton assistance.

Main Methods:

• Acquired EMG signals from four post-stroke patients with hand impairments.
• Trained machine learning models to classify muscle effort based on contraction strength and activation.
• Implemented a closed-loop system to modulate exoskeleton velocity using EMG-based muscle effort.

Main Results:

• The system effectively modulated exoskeleton movements based on EMG readings.
• A reference tracking error of less than 5% was maintained.
• Maximum velocity variation was observed to be 0.7 mm/s.

Conclusions:

• The proposed closed-loop system enables adaptive robotic assistance in stroke rehabilitation.
• EMG-based control allows for personalized and responsive exoskeleton therapy.
• This approach enhances the precision and effectiveness of robotic-assisted neurorehabilitation.