||本次研究計畫分別提出直接與非直接兩種型式的規則調整適應模糊控制系統。簡單的說, 直接型適應模糊控制系統之設計方法是直接從單一的模糊控制器產生控制命令至受控體, 而非直接型適應模糊控制系統之設計方式是先由多個模糊系統對受控體做建模的工作, 然後再依受控體的動態特性產生控制命令而回授至受控體。不論是何種型式的適應模糊控制系統, 其在模糊系統中皆有一能做線上修正之適應定理, 此方式不僅能克服無法做線上尋找最佳參數的缺點, 而且就算受控體有著非常大的不確性, 例如參數的突然改變或外在的干擾等, 也不會受到影響, 因此將使得模糊系統變得更具實用性和強健性。
In this project, we propose direct adaptive fuzzy control system and indirect adaptive fuzzy control system using adjusting the rule base to improve the disadvantages of traditional fuzzy system. Simply speaking, the designed method of direct adaptive fuzzy control system generates a control command from the single fuzzy controller, then feed it back the plant. The designed method of indirect adaptive fuzzy control system is modeling the plant from a number of fuzzy systems at first, then it generates a control command by the dynamic character which will feed it back the plant. Whatever which types of adaptive fuzzy controller, they have the adaptive law of rule regulation on line, respectively. The adaptive law not only can overcome the defect of those approaches which are unable to find the optimal parameters on line, but also the results of the controlled system will not be influenced, even the plant has many uncertainties, such as the suddenly change of parameters or the external disturbance. Therefore, the fuzzy system becomes more practical and robust.