Abstract
本研究是將H_infinity模糊適應控制從狀態可測量的系統擴張到僅輸出可測量的系統。因此，一個觀測器必要裝置在這個適應模糊控制系統來估測系統的狀態。數學模型不確定之非線性系統的追蹤控制本來就相當不易，如果受控的非線性系統又僅輸出可測量，則該非線性系統的追蹤控制(tracking control) 將更形困難。在適應模糊控制技術中，因模糊邏輯系統不可能完全近似非線性系統的不確定部份，故模糊近似僅能視為粗調。在文中，吾人使用H_infinity 技術來做微調，以消除近似誤差與外來干擾對追蹤誤差的影響。本文介紹一個適應輸出濾波器迴路來處理此種H_infinity追蹤問題。 濾波器設計的方法，需要使用某種狀態變數變換。此變數變換的目的，欲增進濾波器對狀態估測誤差的衰減能力。對於f(x)及g(x) 已知的非線性系統，目前輸出回授控制的研究是重要研究潮流之一。對於數學模型不確定又僅輸出可測量之非線性系統的追蹤控制，本文結合適應模糊控制、H_infinity控制和輸出回授控制來解決這個問題。文章對包括倒單擺和機械手臂在內的四個例題進行了電腦模擬，以驗證所提議之以輸出為主的不確定非線性系統之H_infinity適應模糊追蹤控制設計的性能。電腦模擬結果顯示，使用本文所提議的H_infinity適應模糊追蹤控制設計，近似誤差、狀態估測誤差和外來干擾對追蹤誤差的影響，都確實可以被衰減至事先決定好的準位以下。An H_infinity tracking design method via adaptive fuzzy controltechinique fora class of nonlinear systems is proposed to extendfrom state-feedback controlcase to output-feedback control casewhere the state variables are unavailable.The perfect matchingvia adaptive fuzzy logic systems is generally deemed impossible.The H_infinity adaptive fuzzy algorithm is only considered as arough tunning. The effect of approximation error, the estimationerror and external disturbance on tracking error must beattenuated to a prescribed level in order for practical controlapplications.A filtered-output feedback control loop is employedin this work. State transformation is introduced during thedesign process of the filter. Output-feedback control ofnonlinear systems where the system functions f(x) and g(x) areknown attracts a lot of research interest currently. It is thepurpose of this work to combine the adaptive fuzzy control, theoutput-feedback control and the H_infinity control to solve thetracking control problem of unknown nonlinear systems when thestate variables are unavailable. Four examples including thecontrol of an inverted pedulum and the two-link robot arefinally given to illustrate the performance of the proposedmethod. Computer simulation results confirm that the effect ofboth the fuzzy approximation error and external disturbance ontracking error can be attenuated efficiently by the proposedmethod.