在本篇論文中,我們提出一個新的積分可變結構控制器,此控制器結合了小腦模型控制器(CMAC)類神經網路架構和一個平緩的監督控制器,其用於設計未知的非線性系統。由於系統的非線性函數先前並不能夠精確的知道,所以我們利用 CMAC 類神經網路的學習方法去近似未知的非線性函數,同時運用線上即時更新權重因子,完成積分可變結構控制器等效控制的目的。基於李亞普諾夫(Lyapunov)定理,平緩的監督控制器被設計來保證系統的全域穩定。所提出的積分可變結構控制方法經過有效率的學習,不僅減輕了對系統參數的依賴性,並且消除了控制訊號的切跳現象。由於這個以 CMAC 為基礎的積分可變結構控制器被證明是全域的穩定,在所有複雜的訊號是有界的條件下,追蹤軌跡的誤差將會收斂到零。 最後我們運用這個以 CMAC 為基礎的積分可變結構控制器去完成倒單擺系統的控制,並且與適應模糊滑動模式控制器的模擬結果作比較,進而驗證所提出的控制器是同時具有有效性和強健性的成效。
In this thesis, we present a novel integral variable structure control (IVSC) that combines a cerebellar model articulation controller (CMAC) neural network and a soft supervisor controller for use in designing unknown nonlinear system. Since the nonlinear function of system are not known exactly, so using the CMAC neural network learning approach to estimate them, and apply update law to adjust the weighting factors online, and perform the equivalent control on IVSC. Based on the Lyapunov theorem, the soft supervisor controller is designed to guarantee the global stability of the system. The proposed IVSC control scheme alleviates the dependency on system parameters and eliminates the chattering of the control signal through an efficient learning scheme. The CMAC-based IVSC (CIVSC) scheme is proven to be globally stable inasmuch all signals involved are bounded and the tracking error converges to zero. We apply the CIVSC scheme to compare with the adaptive fuzzy sliding mode controller (AFSMC) in an inverted pendulum balancing problem. The simulation results demonstrate the effectiveness and robustness of the proposed controller.