Year

1995

Degree Name

Master of Engineering (Hons.)

Department

Department of Electrical and Computer Engineering

Abstract

A new methodology for design of fuzzy controllers based on Genetic Algorithms has been proposed. The developed design tool initially identifies an approximate model for a system based on a small set of input/output data of the plant. The system identification is also performed based on Genetic Algorithms. The model is then used in the design procedure. The tuning can be carried out either on the membership functions or the fuzzy rules of the fuzzy controller. The method can be applied to linear and non linear SISO and MIMO systems with time delay and unknown structure. Experimental observations show that the developed methodology performs well and is superior to other considered in this study.

Share

COinS
 

Unless otherwise indicated, the views expressed in this thesis are those of the author and do not necessarily represent the views of the University of Wollongong.