摘要
本文提出一种通用的基于模糊聚类和卡尔曼滤波方法的模糊辨识方法．模糊聚类方法在给定的广义目标下按线性簇对被辨识的样本数据进行聚类，这样使得被辨识模型可用若于局部线性模型表示，然后，利用卡尔曼滤波方法拟合这些线性模型．本文给出了详细的模糊辨识算法．为了验证该辨识方法的有效性，本文最后给出了熟知的Ｂｏｘ－Ｊｅｎｋｉｎｓ数据的辨识结果．
This paper discusses a general approach to fuzzy identification based on the fuzzy clusteringtechniques and Kalhian filter method. The fuzzy clustering method utilizes a generalized objective functioninvolving a collection of linear varieties. In this way the identified model is distributed and consists of a series of 'local' linear-type model,then the Kalman filter 'can be used to fit them as accurately as possible. Adetailed identification algorithm is given in this paper. To clarity the advantages of the proposed method,itis used to identify the well-known Box--Jenkins data set,and the result is shown at the end of this paper.
出处
《控制理论与应用》
EI
CAS
CSCD
北大核心
1996年第5期639-643,共5页
Control Theory & Applications
基金
国家自然科学基金
关键词
模糊辨识
模糊聚类
卡尔曼滤波
滤波
fuzzy identification
fuzzy clustering
kalman filter
system identification
