Adaptive Neural Control of a Kinematically Redundant Exoskeleton Robot Using Brain-Machine Interfaces.

Area of Science:

• Robotics
• Neuroscience
• Control Systems

Background:

• Brain-machine interfaces (BMI) offer potential for controlling assistive devices.
• Exoskeleton robots require sophisticated control for effective human-robot interaction.
• Integrating BMI with exoskeletons presents challenges in decoding user intent and ensuring stable control.

Purpose of the Study:

• To develop a closed-loop control system for exoskeleton robots using a brain-machine interface.
• To enable human operators to perform complex manipulation tasks using their electroencephalogram (EEG) signals.
• To achieve optimal and adaptive control of the exoskeleton with minimal prior knowledge of its dynamics.

Main Methods:

• Integration of adaptive controllers in joint space with a velocity-level redundancy resolution method.
• Generation of control commands in task space from BMI-decoded neural activity.
• Implementation of neural dynamics optimization for redundancy resolution and adaptive control for dynamics compensation.

Main Results:

• Successful decoding of human operator's intention from neural activity (EEG).
• Effective control of the exoskeleton robot for manipulation tasks.
• Demonstrated convergence of tracking errors, validating the system's effectiveness and precision.

Conclusions:

• The proposed brain-controlled exoskeleton robot system is effective for manipulation tasks.
• The integrated control strategy provides adaptive dynamics compensation and optimal motion.
• This approach facilitates intuitive and precise human-robot interaction via BMI.