Selecting inference and defuzzification techniques for fuzzy logic control

The aim of this work is to form the basis of a methodology to guide the selection of an appropriate fuzzy logic controller for a given task. Different fuzzy inference techniques used in fuzzy logic controllers are compared initially and then the effects of using different defuzzification techniques on the same fuzzy logic controller are examined. The results of these experiments could then be used as the basis for a selection methodology for choosing an appropriate fuzzy logic controller. An initial examination of the performance of the various fuzzy logic controllers revealed those fuzzy logic controllers that failed to perform satisfactorily. A more detailed comparison, between those different controllers that performed satisfactorily, was made using a wide range of criteria, including the average force applied to the system and the range of initial states of the system that the controller can control.