摘要
讨论了T-S模糊模型的辨识问题,以直线作为数据分类的目标,提出了一种改进的简单辨识算法.首先采用Hough变换,根据给定的输入输出数据,得到了模型后件部分的直线方程,并辨识出结论参数,然后依照得到的直线对输入数据进行分类.考虑输入数据与相应直线的接近程度,以及邻近直线对输入数据的影响程度,辨识出了模型的前件参数.本算法不需要对数据的循环计算,从而大大减少了计算量.仿真例子说明了本算法对T-S模糊模型辨识的有效性.
The identification problem of Takagi-Sugeno fuzzy model is discussed and as an improvement, a simplified method is propo^d taking lines as the objective of data classification. According to a set of given input-output data, linear equations of the consequent of the model are obtained by Hough transform, with conclusive parameters identified. Based on the linear segments, the input data are classified. Considering the proximity of input data to relevant linear segment and how the adjacent linear segment affects the input data, the consequent parameters of the model are identified. The protx)sed method needn' t cyclic data processing, thus reducing greatly the computation. A simulation example shows the effectiveness of the method to T-S fuzzy model identification.
出处
《东北大学学报(自然科学版)》
EI
CAS
CSCD
北大核心
2007年第3期305-307,共3页
Journal of Northeastern University(Natural Science)
基金
国家自然科学基金资助项目(60274099)
教育部流程工业综合自动化重点实验室开放课题资助项目