A neuro-control system for the knee joint position control with quadriceps stimulation

Area of Science:

• Biomedical Engineering
• Neuroscience
• Control Systems

Background:

• Functional Electrical Stimulation (FES) is crucial for restoring movement in individuals with neuromuscular disorders.
• Accurate control of FES-induced movements, like knee joint motion, remains a challenge.
• Existing control strategies often struggle with disturbances and model inaccuracies.

Purpose of the Study:

• To design and evaluate a neuro-PID controller for precise knee joint angle control using quadriceps muscle stimulation.
• To develop an inverse model of the FES-induced quadriceps-lower leg system for feedforward control.
• To assess the controller's efficacy in able-bodied and paraplegic subjects.

Main Methods:

• A multilayer feedforward time-delay neural network was trained as an inverse model for direct feedforward control.
• The Nguyen-Widrow method initialized neural connection weights, and conjugate gradient descent minimized errors.
• A proportional-integral-derivative (PID) controller compensated for residual tracking errors.

Main Results:

• The neural controller achieved precise control of the knee joint angle along a desired trajectory with minimal deviations.
• The PID controller effectively compensated for disturbances and modeling errors.
• The neuro-PID controller demonstrated promising performance in both able-bodied and paraplegic participants.

Conclusions:

• The developed neuro-PID controller is a viable strategy for FES-based knee joint position control.
• This approach holds potential for improving mobility and rehabilitation outcomes for individuals with paralysis.
• Further research can optimize the controller for broader clinical applications.