Fuzzy-neural network controllers using mean-value-based functional reasoning

A new fuzzy reasoning, which is called mean-value-based functional reasoning, is proposed in which the conclusion consists of a function of mean-values on each membership function in the antecedent. From the view point of control theory, i.e. if the conclusion is regarded as a VSS controller, it is possible to rationally design the constant parameters included in the conclusion for the usual functional reasoning and mean-value-based functional reasoning. Furthermore, some fuzzy-neural network controllers are constructed by applying such functional reasonings. Then, the number of learning parameters in the conclusion for the mean-value-based functional reasoning is shown to be reduced drastically, compared with those due to the usual functional reasoning and simplified reasoning. The effectiveness of the proposed method is illustrated by computer simulations for the tracking control problem of a mobile robot driven by two independent wheels.