Identifying Interaction Forces Via EMG Under Changing Motion Dynamics.

Area of Science:

• Biomechanics
• Rehabilitation Engineering
• Human-Computer Interaction

Background:

• Musculoskeletal injuries impede daily activities, necessitating rehabilitation.
• Assistive devices enhance arm mobility during rehabilitation exercises.
• Electromyography (EMG) offers a potential intuitive interface for these devices.

Purpose of the Study:

• Quantify the impact of motion factors on arm muscle activations during dynamic movements.
• Utilize EMG signals and motion factors to detect interaction forces during movement.
• Improve the accuracy of pattern recognition models for unconstrained dynamic movements.

Main Methods:

• Recorded EMG signals during dynamic arm motions with varying positions, forces, and velocities.
• Analyzed changes in EMG features in relation to these motion factors.
• Developed pattern recognition models using EMG signals and motion data (position, velocity) for force prediction.

Main Results:

• EMG features significantly correlate with variations in arm position, interaction force, and motion velocity.
• Pattern recognition models successfully identified intended motion characteristics using only EMG signals.
• Force prediction accuracy improved from 73.77% to 79.17% for elbow flexion-extension with added motion data.

Conclusions:

• EMG signals are sensitive to dynamic motion parameters and interaction forces.
• Integrating EMG with measurable motion data enhances the performance of assistive device control systems.
• This approach holds promise for more intuitive and accurate human-device interaction in rehabilitation.