ILC With High-Order Internal Model For Discrete-Time Systems

This dissertation focuses on the iterative learning control (ILC) design dealing with non-repetitiveness problems such as iteration-varying reference trajectory and unknown parameter which are generated by high-order internal model (HOIM). The main works and contributions are sum-marized as follows.1. By virtue of contraction mapping principle and fixed-point principle, an HOIM-based ILC scheme is proposed for a class of discrete-time linear systems to deal with non-repetitive reference trajectory which is generated from an HOIM. Then a new HOIM-based ILC algo-rithm is generalized to nonlinear systems. Besides, a robust iterative learning controller is proposed to deal with exogenous disturbances and measurement noise, and the convergence condition is given correspondingly.2. An HOIM-based adaptive iterative learning control (AILC) method is proposed for a class of discrete-time nonlinear systems with multiple non-repetitiveness such as unknown parameter which is generated from an HOIM, iteration-varying reference trajectory and initial condi-tion. Asymptotical convergence of the tracking error in the iteration axis is ensured for non-linear systems with unknown control gain and exogenous disturbances by using Recursive-Least-Squares-based learning updating law.3. A projection-based AILC method is proposed to deal with time-iteration-varying parameter which is generated from an HOIM for a class of discrete-time nonlinear systems. Through rigorous proof, we show that the proposed AILC method can deal with random initial condi-tion and iteration-varying reference trajectory, in the sequel achieving asymptotical conver-gence in the iteration domain.4. Considering a class of multiple-input-multiple-output (MIMO) discrete-time nonlinear sys-tems with HOIM-based time-iteration-varying parameter, iteration-varying bounded refer-ence trajectory and random initial condition, a Recursive-Least-Squares-based AILC method and a projection-based AILC method are proposed respectively. The effectiveness of pro-posed algorithm is verified by theoretical proof and simulation.In a word, contraction mapping principle/Lyapunov function based ILC schemes are proposed to tackle the non-repetitiveness for discrete-time nonlinear system under the repeatable environ-ment. The rigorous proofs are presented to solve the non-repetitiveness of known varying law (generated by HOIM) and unknown varying law to show that the system output can converge to the desired trajectory along the iteration axis under the proposed HOIM based ILC method. Simu-lation results demonstrate the effectiveness of the proposed ILC method.