本論文研究目的是針對具有系統耦合特性的多輸入多輸出(MIMO)非線性動態系統，由推導建立雙質量彈簧動態系統輸入與輸出關係的理論模型，且針對非線性提出穩定分析理論。並設計了兩個以模糊理論為基礎的控制器，來使系統能夠同時達到穩定和定位的振動控制。 本論文首先提出模糊自調適PID控制器，運用模糊理論架構簡單的模糊控制器，並藉由可自動調整PID參數機制，讓時變的系統，能夠獲得適時的控制。另一項控制方法是模糊基因演算法控制器，運用改良型基因演算法搜尋調整模糊控制器相關參數，成功改善其演算法在搜尋最佳解問題時，會有不穩定因素、收斂緩慢、不成熟收斂而落入局部最佳解等缺點，穩定收斂且有效搜尋近似最佳解。 經由電腦模擬結果顯示，在雙質量彈簧動態系統中可使之加速穩定收斂，顯示這兩種模糊控制器都具有良好的控制性能，因此以上兩種控制器能夠成功的克服非線性、耦合、參數不定及外在干擾問題。 The purpose of this thesis is focused on the analysis and controller design of non-linear dynamic system with multi-input / multi-output (MIMO) and systematic coupling characteristics. The dynamic model of a two-level mass-spring damper system is first derived; then, two controllers based on the fuzzy theory are developed accordingly. Finally, numerical simulation is carried out to verify the stability and robustness of the controllers. Firstly, a fuzzy-based self-tuning PID controller is proposed based on the fuzzy theory in this thesis and a simple fuzzy controller is structured to adjust PID parameter mechanism automatically; that can let the time-varying system to get in good time control. Another control method is the fuzzy-based genetic algorithm controller and is used the improvement on genetic algorithm to adjust the fuzzy relevant parameter of controller. This control strategy succeeds in improving destabilizing factors, slow-converging, and the local-optimal controller gains with incomplete convergence etc, while algorithms searching the optimal controller gains. It can converge steadily and search effectively the optimal or approximate-optimal controller gains. According to the simulation results of the computer, two-level mass-spring damper dynamic system can make theirs accelerating convergence steadily, and these fuzzy controllers both have good control performance. The above- strategy of two controllers can be successful to solve the programs of the non-linear, coupling, irregular parameter and external interference.
