Robust Sliding Mode Control Based on GA Optimization and CMAC Compensation for Lower Limb Exoskeleton.

Area of Science:

• Robotics and Human-Machine Systems
• Control Engineering
• Biomechanics

Background:

• Lower limb assistive exoskeletons require precise motion intent shadowing to ensure user-exoskeleton coordination.
• Accurate estimation of user intention is crucial for effective position control and seamless collaboration.
• Incoordination can arise if the exoskeleton fails to react appropriately to human motion intent.

Purpose of the Study:

• To propose a hybrid position control scheme for lower limb assistive exoskeletons.
• To enhance the reaction of exoskeletons to human motion intent.
• To improve collaboration between the user and the exoskeleton through precise position control.

Main Methods:

• A hybrid position control scheme combining sliding mode control (SMC) with a cerebellar model articulation controller (CMAC) neural network was developed.
• A genetic algorithm (GA) was employed to optimize the sliding surface and sliding control law for SMC.
• The proposed SMC_GA_CMAC strategy was compared against conventional SMC, SMC_GA, and SMC_CMAC using cosimulation in ADAMS and MATLAB/SIMULINK.

Main Results:

• The proposed SMC_GA_CMAC strategy demonstrated superior position tracking performance compared to other control methods.
• Performance was evaluated in both disturbance-free and bounded disturbance scenarios.
• Cosimulation results confirmed the effectiveness of the hybrid control approach in tracking desired joint angular positions derived from Clinical Gait Analysis (CGA) data.

Conclusions:

• The proposed SMC_GA_CMAC control strategy is effective for controlling lower limb assistive exoskeletons.
• This approach enhances the exoskeleton's ability to react appropriately to human motion intent.
• The developed control strategy can be successfully applied to similar exoskeleton systems for improved functionality and user experience.