System for Automatic Adjustment of Intelligent Controller Parameters

The goal of work is the development of a method for automatically creating a fuzzy controller based on measured data from a control system with a classic controller. To achieve the goal, a mathematical model of the control system in the MatLab Simulink environment has been developed, which allows saving the input and output data of the controller in the modeling process. Analysis of the data series allows determining the parameters of the clusters and the membership functions of the input and output variables of the fuzzy controller. According to the results of clustering, a fuzzy controller rule base can be easily obtained. At the same time, the rule base can have redundant rules not only due to complete duplication, but also due to rules with the same antecedents, but different consequent ones, which leads to uncertainty. To eliminate uncertainty, an algorithm for reducing the rule base has been proposed, based on the selection of redundant rules and the definition of the cluster center corresponding to the new consequent. The authors have developed software that allows to obtain source data from the classical controllers, perform clustering and determination of membership functions parameters, create a rule base, perform its reduction and create a fisstructure for further verification and analysis. The results of modeling a control system with a synthesized fuzzy controller before and after the reduction of the rule base are given.