Autotuning Method of Membership Function in a Fuzzy Learning Controller

Abstract

A learning algorithm for fuzzy inference rules that can tune both the control input and the linguistic membership function of a fuzzy controller is presented. The algorithm introduces a reference model to generate a desired output and minimizes a performance index based on the error between the reference and plant output. To minimize the performance index, the control input and the membership functions are simultaneously updated by a gradient method. Using the algorithm, we can easily achieve good control rules with a minimal amount of prior information about the environment. In particular, when this algorithm is applied to a tracking problem with a nonlinear plant we obtain efficient control rules with fewer number of rules and higher learning speed than those of the fixed membership function.