Research of the Influence of the Sub-Conditions Measure of Significance on the Dynamic Characteristics of a Fuzzy Logic Controller

Abstract

Analysis of operation of the fuzzy control systems shows that there are many factors influencing the quality control such as a number of terms of the input and output variables, forms of values of the linguistic variables constituting its term-sets membership functions, character of the fuzzy relation between the antecedent and consequent spaces (rule base), etc. By varying these factors we can change the parameters of the fuzzy controller for achievement of a satisfactory quality control. Comparing the fuzzy controllers with the traditional linear ones (such as PD-, PID-controllers, etc.) one can notice that the parameters of the linear controllers can be tuned separately. But, when we deal with the fuzzy controllers, any variation of the above mentioned factors has a simultaneous influence on the integral, proportional and differential components of a signal. This is due to numerous logical interactions between the input variables. Furthermore, the specific character of the aggregation of the sub-conditions in the most commonly used fuzzy inference algorithms based on the triangular norm operations does not make it possible to take into account the desired force of each component. An approach proposed in the paper makes it possible to vary one of the signal components within a wide range, leaving the other components practically steady. This in turn simplifies tuning of a fuzzy logic controller.