An efficient self-organizing fuzzy controller using neural networks

Abstract

This paper presents an efficient self-organizing fuzzy logic control scheme using neural networks. Since membership functions regarding change-in-error e/spl dot/ represent the feedback of velocity, they strongly affect transient behaviors of a system. In the paper, therefore, such membership functions are parameterized by the use of the cubic splines. Then, neural networks are adopted to optimize them in self-organizing process. The proposed control scheme consists of a traditional fuzzy logic (FL) controller and a conventional derivative (D) controller. The FL controller is used as a main controller to improve transient behaviors, e.g., a small maximum overshoot and a fast settling time, whereas the D controller is used as an auxiliary controller which is helpful for stabilizing system responses. To demonstrate the effectiveness and the robustness of the proposed method, this self-organizing fuzzy controller is used to control a manipulator. >