sEMG-Based Hand Posture Recognition and Visual Feedback Training for the Forearm Amputee

Area of Science:

• Biomedical Engineering
• Rehabilitation Technology
• Human-Computer Interaction

Background:

• Surface electromyography (sEMG) based gesture recognition is crucial for human-computer interactions, particularly in rehabilitation and prosthetic control.
• High variability in sEMG signals from untrained users often hinders the performance of recognition algorithms.

Purpose of the Study:

• To develop a hand posture recognition algorithm using multichannel sEMG sensors.
• To evaluate the effectiveness of radar plot-based visual feedback training in improving sEMG gesture recognition accuracy.

Main Methods:

• Developed a hand posture recognition algorithm with multichannel sEMG sensors.
• Employed radar plot-based visual feedback for training participants (healthy adults and a bilateral forearm amputee).
• Trained artificial neural network classifiers using single and combined feature vectors.

Main Results:

• Classification accuracy significantly improved in the bilateral forearm amputee after three days of training.
• Visual feedback training effectively reduced sEMG signal variability, enhancing recognition performance.
• The radar plot enabled a bilateral forearm amputee to participate in rehabilitation training.

Conclusions:

• Radar plot-based visual feedback training is an efficient method for improving sEMG-based hand posture recognition.
• This training approach can aid amputees in controlling electric prostheses and participating in rehabilitation.
• The study highlights the potential of visual feedback to overcome sEMG signal variability challenges.