Derivation of Fuzzy Control Rules from Human Operator's Control Actions

Abstract

So called fuzzy control has been developed for the purpose of realization of a man-like controller with the aid of computer. However, most of fuzzy controllers have two very important problems. First is a defect of the reasoning algorithm. Second is a way to acquire control rules, particulaly the precise parameters in the rules. In this Daper considering the above problems, we propose a realistic fuzzy reasoning algorithm and a method to identify control rules from human operators actual control actions. Further we examine the performance of the proposed algorithm by applying it to water cleaning process control.