Hybrid APFPSO Algorithm for Accurate Model-Free Motion Control of a Knee Exoskeleton.

The motor-driven rehabilitation robots can not only simplify the traditional onerous treatment, but also effectively promote the nerve remodeling of patients. The model-free adaptive control algorithm is a data-driven intelligent control algorithm, and it is widely used in rehabilitation robots. However, it faces difficulties with parameter tuning and optimization. This paper proposes an improved particle swarm optimization algorithm to optimize controller parameters, which utilizes tent chaotic mapping to initialize the population and introduces mass into the gravity migration strategy. The original force between two particles is replaced with the attraction and repulsion of the artificial potential field, which calculates the improved perturbation that updates the algorithm’s position and automatically optimizes the control parameters. The experimental results demonstrate that the proposed algorithm can improve the accuracy of trajectory tracking at different amplitude angles compared to the original algorithm, which shows that it gains good potential in motion control of rehabilitation robot.