Development of the human interface equipment based on surface EMG employing channel selection method

Area of Science:

• Biomedical Engineering
• Human-Computer Interaction
• Neuroscience

Background:

• Individual differences in surface electromyogram (SEMG) signals pose challenges for accurate motion recognition.
• Effective control signals are crucial for advanced human-interface equipment.
• Optimizing SEMG measurement channels is key to overcoming inter-subject variability.

Purpose of the Study:

• To develop and validate a novel channel selection method for SEMG-based human-interface control.
• To enhance the accuracy and reliability of hand motion recognition using optimal SEMG measurement channels.
• To reduce the number of required SEMG channels while maintaining high recognition performance.

Main Methods:

• Utilized a 96-channel matrix-type surface electrode array on the forearm to capture SEMG data during hand motions.
• Employed the Monte Carlo method to evaluate 10,000 random channel combinations for optimal selection.
• Selected channels based on achieving the highest motion recognition rate or meeting specific performance thresholds (e.g., >90% recognition).

Main Results:

• The proposed method successfully identified optimal measurement channels for each subject, ranging from 4 to 7 channels.
• Experimental validation with six subjects demonstrated high accuracy in recognizing 18 distinct hand motions, including 10 finger movements.
• Achieved an average real-time recognition rate exceeding 95% for all tested subjects and motions.

Conclusions:

• The developed SEMG channel selection technique significantly improves hand motion recognition accuracy.
• This method offers a subject-specific approach to optimize human-interface control systems.
• The findings suggest a practical and efficient way to implement high-performance SEMG-based interfaces.