Rehabilitation exoskeleton torque control based on PSO-RBFNN optimization.

Area of Science:

• Robotics
• Biomedical Engineering
• Control Systems

Background:

• Exoskeletons are vital in medical rehabilitation but face challenges with control precision.
• Imprecise control can lead to patient accidents, highlighting the need for improved performance.
• Nonlinear dynamics and external interference complicate exoskeleton control.

Purpose of the Study:

• To enhance exoskeleton torque control for safer and more effective patient rehabilitation.
• To address the difficulties in controlling nonlinear exoskeleton models.
• To mitigate interference issues encountered during exoskeleton use.

Main Methods:

• A novel control method combining Particle Swarm Optimization (PSO) and Radial Basis Function Neural Networks (RBFNN) was developed.
• The PSO-RBFNN approach was applied to optimize the torque control of nonlinear exoskeleton joints.
• The study focused on analyzing the motion trajectory and control performance.

Main Results:

• The PSO-RBFNN optimized exoskeleton exhibited faster control speed.
• Enhanced stability, improved accuracy, and stronger anti-interference capabilities were observed.
• The optimized control effectively addressed nonlinearities and external interference.

Conclusions:

• The PSO-RBFNN method significantly improves exoskeleton control performance.
• This optimization enhances safety and efficacy in medical rehabilitation applications.
• The approach provides a robust solution for controlling nonlinear exoskeletons with external interference.