基于标准论域的模糊控制器设计  

Design fuzzy controller based on normalized universe of discourse

作　　者：黄丽[1] 胡世凯[1] 李中夫[1] 李智[1] 

机构地区：[1]四川大学数学学院,成都610064

出　　处：《四川大学学报（自然科学版）》2010年第3期441-446,共6页Journal of Sichuan University(Natural Science Edition)

基　　金：973项目(2002cb312206);国家自然科学基金(10571127);国家自然科学基金(10771147)

摘　　要：设计模糊控制器时常常遇到的困难之一就是为各语言值确定合适的隶属函数.为此本文作者提出了基于标准论域的模糊控制器设计方法.作者首先提出"标准论域"的概念,在标准论域上隶属函数的形式简单、易于处理.然后,通过基础论域到标准论域上的连续单增变换,作者从整体上考虑了一个语言变量所有语言值的隶属函数,而不是像现有的方法一样孤立地确定各个语言值的隶属函数,从而大大减少了工作量.最后,作者详细说明了基于标准论域的模糊控制器的结构,推导了它的输入输出对应关系.现有的模糊控制方法的输入输出关系与基础论域采用的尺度有关,新方法则没有这一缺陷,而且利用这种模糊控制器结构可以方便的实现变论域模糊控制.How to determine proper membership functions of linguistic variables is often extremely important when we design a fuzzy controller, but up to now there not exists an available and general method to deal with this question. This paper establishes a new fuzzy controller based on the concept of ＂normalized universe of discourse＂ （NUOD） and a continuous increase transformation from NUOD to the universe of discourse. This controller is more ease to design than conventional fuzzy controller, and its input-output mapping does not change with the scale of UOD, which is a great limitation of the existing fuzzy controllers. Furthermore, the fuzzy controller of changeable universe of discourse can be carried out conveniently by the structure of fuzzy controller based on NUOD.

关 键 词：模糊控制 隶属函数 标准论域 变论域 

分 类 号：O159[理学—数学] TP273.4[理学—基础数学]