Channel Synergy-based Human-Robot Interface for a Lower Limb Walking Assistance Exoskeleton.

Area of Science:

• Robotics
• Biomedical Engineering
• Neuroscience

Background:

• Surface electromyography (sEMG) enables natural human-robot interaction (HRI), particularly for predicting wearer movement in exoskeleton robots.
• Predicting lower limb movement in paraplegic patients using sEMG is challenging due to weak signals and limited exploration of upper limb sEMG potential.
• Existing HRIs often overlook the synergistic contributions of individual sEMG signal channels.

Purpose of the Study:

• To propose a novel human-exoskeleton interface leveraging upper limb sEMG signals for predicting lower limb movements in paraplegic individuals.
• To introduce a channel synergy-based network (MCSNet) for extracting channel contributions and synergy from sEMG data.
• To validate the effectiveness of the proposed interface and network through experimental data acquisition.

Main Methods:

• Development of a human-exoskeleton interface utilizing upper limb sEMG signals.
• Implementation of a channel synergy-based network (MCSNet) to analyze sEMG feature channels.
• Design and execution of an sEMG data acquisition experiment to test the system's performance.

Main Results:

• The proposed MCSNet demonstrated effective extraction of contribution and synergy from sEMG feature channels.
• The human-exoskeleton interface achieved high accuracy in predicting lower limb movements.
• The system showed robust performance in both within-subject (94.51% accuracy) and cross-subject (80.75% accuracy) scenarios.

Conclusions:

• Upper limb sEMG signals can be effectively utilized to predict lower limb movements in paraplegic patients.
• The MCSNet is a promising approach for analyzing sEMG channel contributions and synergy in HRI.
• This research offers a significant advancement in developing advanced assistive technologies for individuals with paraplegia.