摘要
提出一种基于新的模糊模型和加权递推最小二乘算法 (WRLSA)的非线性系统模糊辨识方法 .新型的具有插值能力的模糊系统可以通过学习从输入输出采样数据中提取MISO系统模糊规则 ,它继承了Sugeno模型及其变化形式的许多优点 .采用相应的模糊隶属函数 ,使得被辨识的模型可用若干局部线性模型来表示 ,然后利用WRLSA拟合这些线性模型 .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 。
A fuzzy identification method for nonlinear systems is suggested based on a new fuzzy model and weighted recursive least square algorithm (WRLSA). The new fuzzy system with interpolating capability extracts fuzzy rules of MISO system from input_output sample data through learning, and inherits many merits from Sugeno_type models and their variations. Through using suitable fuzzy membership function, the identified fuzzy model can be described by several local linear models. And finally, WRLSA is used to fit these linear models. The new fuzzy identification algorithm is proposed. To demonstrate availability of the identification method, the well_known Box_Jenkins data set is also identified.
出处
《控制理论与应用》
EI
CAS
CSCD
北大核心
2003年第1期113-116,共4页
Control Theory & Applications
基金
国家杰出青年基金 (6992 5 3 0 8)
黑龙江省自然科学基金资助项目
