A general technique for deriving analytical structure of fuzzy controllers using arbitrary trapezoidal input fuzzy sets and Zadeh AND operator. (Q1398391)

The author develops a new and more general technique capable of rigorously deriving the analytical structure for any fuzzy controller, of Mamdani type or TS type, as long as it employs the trapezoidal input fuzzy sets and Zadeh fuzzy AND operator. At present, there exists no technique capable of deriving the analytical structure of these fuzzy controllers. The controllers use two arbitrary trapezoidal fuzzy sets for each input variable, four arbitrary singleton output fuzzy sets, four fuzzy rules, the Zadeh fuzzy AND operator, and the centroid defuzzifier. This configuration is more general and complicated than the Mamdani fuzzy PI/PD controllers in the current literature. The author illustrates the technique by applying it to different configurations of the generalized fuzzy PI and PD controllers and shows the derived structures. Regardless of the configurations, the generalized fuzzy PI (or PD) controller is proved to be a nonlinear PI (or PD) controller with variable gain and variable control offset. Even the least general new result is still much more general than the structures reported in the current literature and contains them as special cases.