Aspects of the effects of the defuzzification stage on the stability and the performance of fuzzy model-based control systems

One of the important elements in a fuzzy model-based control system is the defuzzification process. The latter converts the fuzzy control input, which is computed using fuzzy mathematics, into a crisp control input that is fed back to the actual plant. Usually, there is a mismatch between an uncertain system and the fuzzy model that is used to represent it. This may result in difficulties to guarantee certain fundamental properties, such as stability and performance, that the plant must have. This paper discusses the effects of the defuzzification process on the stability and the performance robustness of fuzzy model-based control systems. It outlines the origins of such effects and suggests new directions for further research into this problem.