Tutorial on fuzzy control

Fuzzy logic can be summed up as a technique for handling `lexically imprecise propositions´ and in particular answering queries through propagation of `elastic constraints´. In simpler terms fuzzy logic might be considered as a providing a framework for handling rules (for control or decision making) which have been expressed in an imprecise form. Rule based systems are those which use expertise expressed as a set of rules. Rule based systems need not use fuzzy logic and indeed many have been built which use other forms of knowledge representation. However in control applications, rule based approaches have often needed to work with fuzzy representation because of the way the control expertise has been expressed. One of the key concepts of fuzzy logic is that of a linguistic variable. Take for example temperature which would be a linguistic variable in the control of a heater. It can take on linguistic value such as high, low and quite low. Such linguistic variables are embedded in the rules of a fuzzy controller and allow human control expertise to be represented. The paper considers why fuzzy control can be a technical solution, what the structure of a fuzzy controller is and what applications are being tackled. It concludes by considering where new developments will help in the application of fuzzy control