| Summary: | 碩士 === 元智大學 === 電機工程學系 === 95 === In this thesis, an adaptive backstepping control scheme using fuzzy neural networks, called ABCFNN, is proposed for a class of nonlinear uncertain nonaffine systems. The nonlinear nonaffine system contains of external disturbance, uncertainty, or parameters variations. Two kinds of fuzzy neural network systems (FNNs) are used to estimate the unknown system functions. According to the estimations of the FNNs, the ABCFNN control input can be chosen by backstepping design procedure such that the system output follows the desired trajectory. Based on the Lyapunov approach, the adaptive laws of FNNs’ parameters are obtained. To solve the effect of FNNs’ membership functions initialization, the back-propagation algorithm and Taylor expansion method are adopted to derive the update laws of FNNs’ parameters m, ??, and θ. Besides, the proposed ABCFNN is extended to a class of nonlinear cascade systems. Finally, the proposed ABCFNN is applied to the controlling of a CSTR system and a single-link flexible-joint robot. Simulation results are shown to demonstrate the performances of our approach.
|
|---|
