於本博士論文中，分別針對具未知項、雜訊項或時間延遲的 Takagi-Sugeno 模糊系統，提出一些觀測器設計法。在第二章，針對具未知項與雜訊項之 T-S 模糊系統的觀測器設計已被提出。其中，根據系統狀態與輸出雜訊的整合，增廣型 T-S 模糊系統已被建立。接著基於 Lyapunov 理論及線性矩陣不等式工具，模糊觀測器及其合成的充分條件已被實現，使得系統狀態與輸出雜訊可同時被估測。於數值範例中，將展示所提出的觀測器能夠保證期望目標。與第二章有相同的系統與相似的控制程序，將於第三章提出一種適應性觀測器，實現不具額外項 的狀態與雜訊估測。本章也提出兩個實際範例與一些文獻比較。因為存在不可利用的狀態，所以第四章提出基於觀測型之 H∞ 模糊控制器，以保證不具雜訊之 T-S 模糊系統能夠被漸進估測，或具雜訊之T-S 模糊系統能夠滿足 H∞ 控制效能。於本章節中，在狀態系統與觀測器間，我們不需要此項 所以觀測器將比之前變得更簡單。兩個實際範例將顯示所提出之觀測型 H∞ 控制器的有效性。在第五章，於第二章中所介紹之觀測器方法將被應用於離散型具時延 T-S 模糊系統。狀態與雜訊也能夠被所設計之觀測器估測。In the dissertation, several observer design methods for Takagi-Sugeno (T-S) fuzzy systems with uncertainty, disturbance or/and time-delay are proposed, respectively. In Chapter 2, the observer design for T-S fuzzy systems with uncertainty and disturbance is presented, where an augmented T-S fuzzy system is constructed by integrating the system state and the output disturbance into a new variable. Then, based on Lyapunov theory and LMI tools, some sufficient conditions for the existence of the fuzzy observer and the observer synthesis are achieved such that the system state and the output disturbance are estimated simultaneously. A numerical example demonstrates the proposed observer can guarantee the desire goal. In Chapter 3, for the same system and with similar design process in Chapter 2, an adaptive observer is proposed to achieve the state and disturbance estimation without extra term . This chapter also proposes two practical examples and gives some comparison with other related paper. Because of the existence of unavailable states, Chapter 4 proposes a fuzzy observer-based H∞ controller to guarantee the asymptotical estimation for the T-S fuzzy system without disturbance and H∞ control performance for the system with disturbance. In this chapter, we do not need the term in both of state system and observer so that the observer becomes simpler than before. Two real examples are illustrated to demonstrate the effectiveness of the fuzzy observer-based H∞ controller. In Chapter 5, the observer design method in Chapter 2 is applied to discrete time-delay T-S fuzzy systems. The state and disturbance are estimated by the designed observer too.
