Indirect Adaptive Self-structuring Fuzzy Neural Network Control System

Abstract

To obtain a high control performance of adaptive fuzzy control systems, the stability guarantee should be made stricter to reduce the approximation error of the system. Avoiding the system becomes complex, this study aims to design a kind of adaptive fuzzy control system that realizes compatibility of concise stability guarantee and high control performance, and it proposes the adaptive fuzzy control system design method of the indirect controller approximation type designed based on the self-structuring fuzzy neural network. It is evaded that the approximation error element that was not able to be removed by adaptive law influences the entire control input directly, and influences are distributed, by composing an indirect type control system. Additionally, it aims at the improvement of the control performance by compensating this remaining approximation error by the robust control input element in the control input. Finally, the effectiveness of this method is verified by the computation simulation of the cart-pole system control.