Study On Adaptive Iterative Learning Control Algorithm And Application

In practical systems, considering the existence of various nonlinear, parameter uncertainty and actuator faults, the conventional model-based control methods have difficulty in achieving the high-precision tracking. Iterative learning control(ILC) is a tracking control method mainly applied in the repetitive motion systems. ILC doesn’t depend on the system model information and the control input only contains previous control and current tracking error. Owing to the defaults of ILC in dealing with the uncertain problems, the adaptive control is added into the ILC control called adaptive iterative learning control(AILC). AILC has the double advantages of ILC in solving the problem of repeated tracking and adaptive control in solving the problem of system uncertainties.The main works and contributions are summarized as following:① An adaptive iterative learning control algorithm under alignment condition is proposed for tracking control designs for robot manipulators with considering of modelling uncertainties and external disturbances. The proposed controller contains three parts: PD-type feedback part, parameter estimation part and external disturbance compensation part. The convergence analysis of proposed control scheme is based on Lyapunov composite energy function in the iteration domain containing position tracking error, velocity tracking error and parameter estimation error along both the time and iteration axis. Numerical simulations also confirm and verify the effectiveness of the proposed method.② Consider the tracking control in Automatic Train Control and the repetitive motion characteristic in high-speed train system, a fault-tolerant adaptive iterative learning control algorithm is proposed for tracking control in high-speed train operation system considering the factors of actuator faults, speed delay, parameter uncertainty and control input saturation. The algorithm contains an actuator efficiency factor for system fault model without needs for precise model information nor fault occurrence time constant concurrently. By Lyapunov-Krasovskii composite energy function, the convergence analysis of proposed algorithm is given. The comparison results include that the proposed algorithm has faster convergence speed, and the convergence error has a good transient performance in iteration domain.By comparing with the conventional iterative learning control, more uncertainties are concurrently considered when the controllers are designed, including modelling uncertainties, external disturbances, parameter uncertainty, control input saturation, speed delay, and so on. Meanwhile, as to the application, applying the iterative learning control method automatic train control area effectively deals with the existent problems under those control methods which ignore the fact of repetitive motion characteristic in high-speed train system.