一种能同时获得模型阶次和参数的递推辨识算法
A New Recursive Algorithm for Simultaneous Identification of Model Order and Parameters
-
摘要: 对SISO线性差分方程模型,可以构造一种新的数据乘积矩矩阵,称信息压缩阵.对信息 压缩阵进行递推UD分解,便得到一种新的递推辨识算法,它能在每步递推中同时得到模型 的阶次和参数,从而降低了辨识过程的计算量.该方法还同时具有良好的数值计算品质.Abstract: For linear difference equation model with unknown order, a new covariance matrix call-ed "the Condensed Information Matrix" (CIM) can be constructed. Applying UD factorization technique to CIM, a new recursive identification algorithm can be obtained. The new algorithm can simultaneously identify the model order and parameters, so that the computational burden can be reduced.
计量
- 文章访问数: 2250
- HTML全文浏览量: 58
- PDF下载量: 982
- 被引次数: 0