【摘要】 ＲＢＦ网络是一种广泛应用的神经网络模型，而Ｔａｋａｇｉ－Ｓｕｇｅｎｏ模糊推理规则是一种简化的模糊推理规则，两种方法的起源不同．文中分析了在一定条件下，ＲＢＦ网络与简化的Ｔａｋａｇｉ－Ｓｕｇｅｎｏ模糊推理规则的函数等效性，揭示了网络权值与推理规则参数的对应关系，从而为两种方法的互换使用奠定了理论基础．在此基础上，提出了使用ＲＢＦ网络在实时控制过程中为一些复杂的被控对象获取Ｔａｋａｇｉ－Ｓｕｇｅｎｏ型模糊控制器参数的方法．以ＲＢＦ网络作为控制器，在实时闭环控制过程中在线地修正网络权值，通过有限次的学习即可获得较优的网络权值，根据对应关系获得Ｔａｋａｇｉ－Ｓｕｇｅｎｏ型模糊控制器的规则参数．一混流式水轮机组的仿真结果证明了该方法的有效性．更多还原

【Abstract】 RBF network and Takagi Sugeno fuzzy inference rule are two methods with different origins. The former is one of the widely used neural network models, and the latter is a kind of simplified fuzzy inference rules. The paper shows the function equivalence under certain conditions between the two methods and the corresponding relationship between the RBFN weights and the fuzzy inference rule parameters, which lays the foundation for mutual exchange of the two methods in some applications. On the basis of it, a new approach is proposed, in which RBFN works as a controller to obtain Takagi Sugeno fuzzy controller parameters for some complex objects online. The RBFN weights are modified in the course of control process. After learning, the proper values of RBFN weights are obtained,and accordingly the Takagi Sugeno controller parameters are also obtained. The simulation results prove the validity.更多还原