Processing Surface EMG Signals for Exoskeleton Motion Control

Area of Science:

• Biomedical Engineering
• Robotics
• Neuroscience

Background:

• Surface electromyography (sEMG) signals are used for controlling robotic assistive devices.
• Challenges remain in improving system accuracy and signal processing for noise reduction.

Purpose of the Study:

• To present procedures and validation for EMG-driven speed-control of an exoskeleton and integrated treadmill.
• To enhance user-system interaction for robotic assistive ambulation.

Main Methods:

• Extracted gait cycle duration (GCD) from sEMG using autocorrelation and Bayesian fusion algorithms.
• Programmed GCDs to control exoskeleton motion speed.
• Validated the system with 6 healthy volunteers.

Main Results:

• Autocorrelation algorithm extracted GCD from individual muscle contractions with variability.
• Bayesian fusion processed multiple muscle GCDs to yield a final GCD with minimal variance.
• Fused GCD effectively controlled exoskeleton and treadmill speeds, showing higher EMG amplitude and shorter GCD at faster walking speeds.
• Exoskeleton joint motion tracks matched algorithm-generated trajectories with high agreement.

Conclusions:

• EMG-driven speed-control offers intuitive human-machine inter-limb coordination for robotic assistive systems.
• The system shows potential for gait rehabilitation in patients with incomplete spinal cord injuries.