模糊规则在线提取及其在非线性系统辨识中的应用  

On-Ling Extraction of Fuzzy Rule and Its Application in Non-Linear System Recognition

作　　者：黄自元[1] 李力雄[1] 费敏锐[1] 

机构地区：[1]上海大学机电工程及自动化学院,上海200072

出　　处：《自动化仪表》2001年第10期10-11,16,共3页Process Automation Instrumentation

基　　金：教育部高等学校骨干教师资助计划项目;上海市教委曙光计划项目

摘　　要：提出一种从输入 -输出数据提取模糊规则的新方法 :基于K Nearest Neighbor概念对输入 -输出数据对进行预处理 ;利用竞争学习对输入空间自适应聚类 ;提取高斯基 ;在线调整规则库和模糊推理。将该方法应用于非线性系统辨识 。A new method of extracting fuzzy rule from input and output data is staed. Based on K-Nearest-Neighbor concept the pre-processing on input/output data pair is realized; the self-adaptive category gathering is operated on input space by using competition learning; gaussage is extracted; the rule library and fuzzy inference are adjusted on line. The method has been used on non linear system recognition, the simulation resuit shows its effectiveness.

关 键 词：模糊规则 模糊聚类 K-Nearest-Neighbor 非线性系统辨识 

分 类 号：O231[理学—运筹学与控制论]