Adaptive Iterative Learning Control of Nonlinear Systems Using an Output Recurrent Fuzzy Neural Network

• Main Authors: Ying-Chung Wang, 王盈中
• Other Authors: Ching-Cheng Teng
• Format: Others
• Language: en_US
• Published: 2003
• Online Access: http://ndltd.ncl.edu.tw/handle/75853503512576087572

博士 === 國立交通大學 === 電機與控制工程系 === 92 === In this thesis, we propose an adaptive iterative learning control (AILC) based on a new output recurrent fuzzy neural network (ORFNN) for solving the traditional iterative learning control (ILC) problems. The proposed AILC will be classified into indirect AILC (IAILC) and direct AILC (DAILC) categories. A systematic design procedure of the adaptive iterative learning controller is given for a class of repeatable nonlinear systems with unknown nonlinearities and varying initial resetting errors. Furthermore, the complete analysis of stability and convergence is also furnished. The proposed IAILC is first presented to a class of unknown nonlinear plants with initial varying state errors. In order to overcome the design difficulty due to initial state errors at the beginning of each iteration, a concept of time-varying boundary layer is employed to construct an error equation. The learning controller in the IAILC is designed by using the given ORFNN to approximate the plant's nonlinearities. Some auxiliary control components are applied to eliminate approximation error and ensure learning convergence. Since the optimal ORFNN parameters for a minimum approximation error are unavailable, an adaptive algorithm with projection mechanism is derived to update all the consequent, premise and recurrent parameters during iteration processes. Only one ORFNN is required to design the proposed IAILC. Based on a Lyapunov like analysis, we show that all adjustable parameters and internal signals remain bounded for each iteration. Furthermore, the norm of state tracking error vector will asymptotically converge to a tunable residual set as iteration goes to infinity. On the other hand, the proposed DAILC is then applied to the same problem which deals with the iterative learning control of nonlinear plants with varying initial resetting errors and unknown plant's input gain. Since the optimal equivalent controller is unknown, the ORFNN is used to approximate the certainty equivalent controller, but not to model the plant nonlinearities. It is shown that the stability, convergence as well as the learning performance are still guaranteed similar to those results given in the IAILC case.