AN EXPERIMENTAL INVESTIGATION OF THE DYNAMIC PERFORMANCE OF FUZZY-LOGIC CONTROL

This paper presents the results of an experimental (simulation based) investigation of the effect of scaling factor choice on the step response of a fuzzy logic controlled system. It is shown that if scaling factors are adjusted to give the best step response than small changes of process parameters (eg, time constant and gain) quickly degrade this good performance. Alternatively, if the performance required is degraded from the best step response obtainable by tuning the scaling factors, then scaling factors can be selected to give an acceptable performance over a wide variation of process parameters.