Knowledge-based Adaptation of Controllers

Abstract

The design of PID controllers, identification of theirs parameters and finally the control quality monitoring are problems often discussed in many applications. Many of the methods have been developed, classic or not so common methods using approaches of artificial intelligence. This thesis is focused on the design of an intelligent controller. All parts of the proposed design are based on non-conventional methods which are very efficient even in such typical engineering field, as the system's identification, monitoring, and designing of PID controller's parameters. The identification of system's parameters is based on optimization of parameters of its difference equation using genetic algorithms. The continuous monitoring of control process quality is done using a fuzzy expert system of Mamdani type. For designing of~parameters of PID controller is also used fuzzy expert system, but of Takagi-Sugeno type. The evaluation of the proposed methods and numerical experiments are presented (simulated) using the software environment Matlab and Matlab&Simulink.