Hand Gestures Recognition for Human-Machine Interfaces: A Low-Power Bio-Inspired Armband

Area of Science:

• Biomedical Engineering
• Human-Machine Interface (HMI)
• Wearable Technology

Background:

• Hand gesture recognition is increasingly vital in biomedical Human-Machine Interfaces (HMIs).
• Non-invasive techniques like surface Electromyography (sEMG) and PhotoPlethysmoGraphy (PPG) are commonly used.
• Growing interest from academia and industry fuels the development of wearable devices for diverse applications.

Purpose of the Study:

• To develop a novel 7-channel sEMG armband for HMI applications.
• To enable on-board computation of the Average Threshold Crossing (ATC) parameter for gesture recognition.
• To create a low-power, efficient wearable device for gaming and rehabilitation.

Main Methods:

• Designed and prototyped a 7-channel sEMG armband.
• Implemented on-board computation of the Average Threshold Crossing (ATC) parameter.
• Acquired sEMG data from 26 participants performing hand gestures.
• Trained and evaluated a real-time gesture recognition classifier.

Main Results:

• Achieved an average classifier accuracy of 91.9% for recognizing 8 hand gestures and an idle state.
• Demonstrated low power consumption (2.92 mA) and prediction latency (1.34 ms).
• The device is capable of long-term operation (up to 60 hours).

Conclusions:

• The novel sEMG armband offers an efficient and accurate solution for hand gesture recognition.
• Its low-power design makes it suitable for extended use in medical and consumer applications.
• The on-board ATC computation enhances the device's power efficiency and real-time capabilities.