EMG Dataset for Gesture Recognition with Arm Translation.

Area of Science:

• Biomedical Engineering
• Robotics
• Human-Computer Interaction

Background:

• Myoelectric control systems are vital for prosthetics and robotics.
• System accuracy is hindered by factors like arm position changes.
• The impact of arm position on myoelectric signal quality is understudied.

Purpose of the Study:

• To address the gap in understanding arm position effects on myoelectric control.
• To introduce a novel dataset of surface electromyographic (EMG) signals across varied arm positions.
• To facilitate the development of position-invariant myoelectric control algorithms.

Main Methods:

• Collected surface EMG and hand kinematics data from 8 participants.
• Participants performed 6 distinct hand gestures across multiple arm positions.
• Developed a novel data acquisition protocol for future EMG data collection.

Main Results:

• A comprehensive dataset capturing EMG signal variability with arm position is now available.
• The dataset enables investigation into position-invariant myoelectric control decoding.
• The proposed protocol supports future data collection efforts.

Conclusions:

• The novel dataset is a valuable resource for training and benchmarking position-invariant myoelectric control algorithms.
• Addressing arm position variability is crucial for enhancing myoelectric control robustness.
• The study promotes advancements in prosthetic limb and robotic control.