Wearable high-density EMG sleeve for complex hand gesture classification and continuous joint angle estimation.

Area of Science:

• Biomedical Engineering
• Human-Computer Interaction
• Neuroscience

Background:

• High-density electromyography (HD-EMG) offers a natural interface for human-computer interaction (HCI).
• Decoding muscle activity for complex hand movements remains a challenge in EMG-driven HCI.

Purpose of the Study:

• To demonstrate the capability of a novel HD-EMG forearm sleeve for high-resolution muscle activity capture.
• To decode complex hand gestures and estimate continuous hand position using joint angle predictions.
• To evaluate algorithm modifications for minimizing decoder latency and reducing training requirements.

Main Methods:

• A 150-electrode HD-EMG forearm sleeve recorded muscle activity from ten participants performing 37 hand movements.
• Motion capture data from 23 joint angles were used for training regression models.
• Decoding algorithms were evaluated for gesture classification and continuous joint angle estimation.

Main Results:

• The HD-EMG sleeve achieved high accuracy in classifying hand gestures (up to 90% in sequential sets) and estimating joint angles (R² up to 0.88).
• Median absolute error for joint angle estimation remained below 10° across all joints.
• Algorithm modifications improved decoder performance by minimizing latency and reducing training data needs.

Conclusions:

• The HD-EMG sleeve, coupled with advanced machine learning, provides a powerful tool for hand gesture recognition and joint angle estimation.
• This technology holds significant potential for applications in prosthetics, assistive technology, rehabilitation, and human-robot collaboration.