基于Fisher投影和变论域思想的模糊控制器降维方法及其在液位问题的应用

Dimension Reduction of Fuzzy Controller Based on Fisher Projection and Variable Universe Method and Its Application on Water Level Control

维数问题是模糊控制器领域的一个重要问题。本文结合Fisher降维、ANFIS模糊神经网络和变论域控制思想,论述了一种可以减少输入个数的、高精度的模糊控制器的构造方法。以液位控制实验为例,对原有的控制器进行了降维,发现这种方法可以极大地减少了控制器的计算量,同时提高控制器控制品质,具有较高的推广价值。

Dimension explosion problem is a hot issue in fuzzy control. This paper proposes a method to reduce the number of input variables of a fuzzy controller based on Fisher Projection method, Artificial Network Fuzzy Inference System and Variable Universe Control Method. The experiment in the problem of water level control shows that this method can greatly improve both the compute efficiency, and the control quality. It can be applied to similar control problems.