Analysis and Synthesis Methods for Fuzzy Logic Controllers

Fuzzy logic in control has been successful used to capture heuristic control laws obtained from human experience or engineering practice in automated algorithm. These control laws are defined by means of linguistic rule, for example ``if the pressure is high, then decrease the pump power''. The heuristic approach in the controller design can be appealing for its simplicity, but formal design method can be mandatory in some cases.This thesis aims at defining some original method for the analysis and design of fuzzy logic controller. More in detail some properties of Variable Structure Control theory have been used, in particular the sliding mode dynamics, to design the rules and the membership functions of a Fuzzy Logic Control. In the second part of the thesis a model-based procedure for the synthesis of a fuzzy logic controller is presented. This algorithm defines the linguistic rules of a fuzzy logic controller, starting from a desired closed loop behavior and a plant model, both described via linguistic statements. To obtain a plant model in a linguistic form, suitable for the above algorithm, a novel identification procedure is proposed.