SYNTHESIS OF A FUZZY SUPERVISOR, ADJUSTMENT OF PID CONTROLLER COEFFICIENTS BASED ON INFORMATION ABOUT THE ERROR AND ITS INPUT

Abstract

The article considers the method of synthesis of a fuzzy supervisor, corrections of coefficients of the PID controller based on information about the error and its derivative. The method of the algorithm of adaptation of the reaction of the controlled object to external disturbances based on the estimation of the control error and its derivative using a fuzzy supervisor is considered. The idea of using a fuzzy supervisor is to organize a two-level system in which the lower level is a conventional PID controller, and the upper level is a fuzzy controller. The fuzzy supervisor is connected in parallel with the classical PID controller. The use of a fuzzy supervisor allows for adaptive adjustment of fixed transfer coefficients of the PID controller over the entire transient process, which allows significantly increasing the robustness of the control system. The increase in robustness is achieved by the fact that fuzzy controllers, which are inherently nonlinear, can control linear objects better than classical controllers, and even control nonlinear objects for which linear controllers are unsuitable.Classical control theory is mainly focused on the synthesis of linear controllers for such objects. The property of linearity allows us to obtain convenient algorithms for estimating such important system properties as stability, controllability, observability, etc.However, all real objects are nonlinear, and the linearization of their mathematical description requires consideration of a certain operating point, relative to which small deviations are considered.Based on the results of the synthesis of a fuzzy supervisor, tables for setting linguistic rules for changing the transfer coefficients of the PID controller were constructed, and a graphical representation of the control surfaces of the supervisory system was obtained. The acquired graphs of the transient process of the control object when using a classical PID controller and a fuzzy supervisor, corrections of the coefficients of the PID controller were compared.