A new methodology of fuzzy constraint-based controller design via constraint-network processing

Abstract

A complete design framework for a fuzzy constraint-based controller based on fuzzy-constraint processing and its semantics and relationship to fuzzy logic is presented. In this paper, the concept of "fuzzy constraints" in problem solving is introduced, and some basic definitions of fuzzy-constraint processing in a constraint network and its semantic modeling are addressed. Then a fuzzy local propagation inference mechanism for reasoning about imprecise information applying the filter operation in a network of constraints is proposed. Moreover, we advance the concurrent fuzzy-logic controller (FLC) to a new type of controller, the fuzzy constraint-based controller (FCC), using a more general predicate calculus and full first-order logic knowledge representation and making use of the idea of fuzzy-constraint processing to model practical dynamic control systems. Finally, simulation results show that a FCC achieves equivalent performance as PD type and PI type FLCs and it also demonstrates superior outcomes to a conventional PID controller in terms of rise time and peak-percent overshoot.