An Exosuit System With Bidirectional Hand Support for Bilateral Assistance Based on Dynamic Gesture Recognition

Area of Science:

• Robotics
• Biomedical Engineering
• Artificial Intelligence

Background:

• Hand motor impairment significantly impacts the daily lives and independence of the elderly.
• Existing assistive devices often lack sophisticated control for nuanced hand movements and bilateral coordination.

Purpose of the Study:

• To develop and evaluate an electromyography (EMG) exosuit system with bidirectional hand support for bilateral coordination assistance.
• To enhance daily task performance for individuals with hand motor impairments using an intelligent assistive device.

Main Methods:

• Developed a novel exosuit system comprising hardware (exosuit jacket, backpack, EMG module, bidirectional glove) and software.
• Implemented a dynamic gesture recognition model utilizing Graph Convolutional Network (GCN) and Long Short-Term Memory (LSTM) networks for EMG signal analysis.
• Conducted offline training and online control experiments to validate gesture recognition and system effectiveness.

Main Results:

• Achieved a high gesture recognition rate of 96.42% ± 3.26%, outperforming traditional models.
• Demonstrated successful completion of daily tasks by all subjects with high bilateral coordination assistance success rates (88.75% and 86.88%).
• Verified the exosuit system's effectiveness in providing bidirectional hand support for daily activities.

Conclusions:

• The developed EMG exosuit system effectively assists elderly individuals with hand motor impairment in performing daily tasks through bidirectional hand support and bilateral coordination.
• The proposed GCN-LSTM based dynamic gesture recognition model offers a superior approach for controlling assistive devices.
• The system and methodology show promise for broader applications in various limb assistance scenarios.