Synthesis of the Neuro-Fuzzy Adaptive Control System of a Dynamic Object

Abstract

In this scientific work, the method for creating an adaptive control system of the discrete dynamic object, which contains a compensator, emulator, and controller, based on the Sugeno fuzzy model and allowing eliminating the influence of disturbances and control error is purposed. Algorithms for parametric and structural identification are synthesized, which, along with the method of back propagation of errors, were used to adapt neuro-fuzzy models. The proposed hybrid models based on neural networks and fuzzy logic allow ensuring the high process efficiency and controlling quality.