A Low-Cost, Wireless, 3-D-Printed Custom Armband for sEMG Hand Gesture Recognition

Area of Science:

• Biomedical Engineering
• Wearable Technology
• Machine Learning

Background:

• Wearable technology enhances human capabilities for complex tasks.
• Surface electromyography (sEMG) armbands derive human intent using machine learning.
• Existing sEMG systems are often costly or compromise quality/wearability.

Purpose of the Study:

• Introduce the low-cost, 3DC Armband for sEMG acquisition.
• Compare the 3DC Armband's performance against the popular Myo Armband.
• Provide open-access data and design resources for the 3DC Armband.

Main Methods:

• Developed a wireless, 10-channel, dry-electrode 3DC Armband with an IMU.
• Collected an offline dataset of 22 participants performing 11 hand/wrist gestures.
• Simultaneously recorded data from the 3DC Armband and Myo Armband.
• Utilized three classifiers with three input modalities for performance evaluation.

Main Results:

• The 3DC Armband significantly outperformed the Myo Armband (p < 0.05).
• Superior performance was observed with training data of 10 seconds or more per gesture.
• The 3DC Armband demonstrated robust performance across different classifiers and input modalities.

Conclusions:

• The 3DC Armband offers a cost-effective and high-quality alternative for sEMG acquisition.
• The developed system and dataset facilitate further research in human intent recognition.
• Open-access resources promote wider adoption and development of myoelectric control systems.