Type-2 Neuro-Fuzzy Control for a Class of Nonlinear Systems
Shixi Hou, Cheng Wang, Suwei Zhai, Yundi Chu · 2021
In this paper, a type-2 neuro-fuzzy control for a class of nonlinear systems is studied. Firstly, an integral-type sliding mode controller (SMC) is designed to ensure that the error converges in a finite time. At the same time, the saturation function as an effective way to alleviate chattering is utilized. Moreover, a type-2 neuro-fuzzy networks (T2NFN), in which network parameters can be adjusted online, is used to approximate the designed SMC. In order to improve the generalization ability, T2NFN combines a recursive feature selection algorithm. In particular, due to the added robust compensator, the issue of the approximation error also can be overcome. Finally, the T2NFN controller is applied to the active power filter (APF) to show its superiority.