一种用CMAC网络建立模糊控制器的新方法

A new method to design fuzzy controller with CMAC network

提出了一种新的建立模糊控制的方法，在原来模糊控制的基础上，对模糊集提出了新的解释．将模糊集的建立和模糊规则的确立和受控系统的物理行为直接联系起来．然后，通过ＣＭＡＣ神经网络对受控系统进行学习，掌握受控系统的特性，并最终由ＣＭＡＣ神经网络根据对模糊集的新的解释，来建立模糊集并提取模糊规则．这样，使模糊控制的建立避免了由于设计者而带来的偶然性．

In this paper, they present a new method to design fuzzy controller. This paper sets up a new explanation for fuzzy set and fuzzy rule based on imal fuzzy controller.Under this explanation, relation has been set up between the setting of fuzzy set, fuzzy rule and the physical action of the controller's object. Then, by using the CMAC network to learn the action of the controller's object,designer can get the knowledge of the object. After that, under the new explanation for fuzzy set and fuzzy rule, the fuzzy controller can be designed automatically. So this method avoids the casuality caused by the designer. The self learning ability of CMAC also makes it possible for on-line self adaptive fuzzy controller.