利用並聯控制器架構穩定非線性非最小相位系統

Abstract

在本論文中,我們提出一個並聯的控制器架構,以穩定非線性單一輸入 單一輸出非最小相位系統.這個架構由一個傳統的控制器和一個模糊類神 經網路控制器組合而成.傳統控制器是用來穩定系統的主要部份.它的推導 方式是先對系統做部份迴授線性化,使系統轉換成為簡單的串接形式.然 後,我們再以 Lyapunov 的觀念為基礎,利用串接控制設計技巧去分析這個 系統的控制法則.至於模糊類神經網路控制器的作用則是適應性地補償傳 統控制器的不足.它可以說是一個微調器.藉由模糊類神經網路控制器的幫 助,這個並聯的控制器可以改善系統的響應性能,並且對當有外力干擾或有 較大初始值時,有好的容忍性.最後,對例子做電腦模擬的結果,顯示了這個 並聯控制器有好的成效和實用性. A parallel control approach is presented in this thesis to stabilize nonlinear single-input-single-output systems which are of nonminimum phase.The architecture consists of a conventional controller and a fuzzy neuralnetwork controller. The conventional controller is the main part for stabilization. It is derived first using partial feedback linearizationwhich transforms the system into the simplified cascade connection form.Then we apply the cascade control design technique based on the notion ofLyapunov to analyze control laws for the system. The fuzzy neural networkcontroller adaptively compensates for the conventional controller. It actsas a fine tuner. With the aid of the fuzzy neural network controller, theparallel controller can improve the performance of the system, has abilityof disturbance rejection and can enlarge the domain of attraction forlarger initial conditions. Furthermore, computer simulations of examples are presented to show the performance and applicability of this proposedcontroller.