Closed loop iterative learning control for consistency tracking in lower limb rehabilitation robotic system with initial state deviations

In the research on consensus tracking control for Lower Limb Rehabilitation Robotic Systems (LLRRS), it is crucial to ensure that all state variables of the LLRRS, including initial state, angle, and angular velocity, converge towards a consensus. This paper addresses the motion tracking control issue of LLRRS in scenarios with initial state deviations. Firstly, a dynamic mathematical model of the LLRRS is established, and the target motion trajectory is determined. To tackle the challenges posed by initial state deviation, a closed-loop PD-type accelerated iterative learning controller with initial state learning is designed, utilizing only the output measurements of the system and a variable learning gain factor. The applicability of this controller for achieving consensus tracking control of the LLRRS state variables is verified through mathematical analysis and simulation. Finally, the feasibility and effectiveness of the proposed algorithm are corroborated through experimental prototype testing. The experimental results demonstrate that the maximum tracking error for the hip joint angle of the LLRRS is 7.14°, and the maximum tracking error for the knee joint angle is 5.74°.