Real-Time Hand Motion Recognition Using sEMG Patterns Classification.

Area of Science:

• Rehabilitation Engineering
• Biomedical Signal Processing
• Human-Computer Interaction

Background:

• Current prosthetic hands require more intuitive and effective control strategies.
• Reducing the cost of myoelectric prostheses while increasing performance is a key goal.

Purpose of the Study:

• To present a real-time approach for classifying finger motions using surface electromyography (sEMG) signals.
• To develop an accessible and effective control system for prosthetic hands.

Main Methods:

• Utilized a 7-channel multichannel signal acquisition platform built with off-the-shelf components.
• Implemented real-time sEMG pattern classification using Linear Discriminant Analysis (LDA).

Main Results:

• Successfully identified thirteen distinct hand motions.
• Achieved up to 95.8% accuracy, with an average accuracy of 92.7% across 8 participants.
• Provided updated predictions every 192 milliseconds.

Conclusions:

• The developed real-time sEMG classification system offers a promising, accurate, and efficient method for prosthetic hand control.
• This approach can significantly enhance the intuitiveness and performance of myoelectric prostheses, benefiting rehabilitation engineering.
• The use of off-the-shelf components suggests a cost-effective solution for improving prosthetic technology.