递推最小二乘法中数据饱和现象的一种消除方法

ONE METHOD OF ELIMINATING THE "DATA SATURATION" PHENOMENON IN THE RECURSIVE LEAST SQUARES METHOD

递推最小二乘法是经常使用的一种参数估计方法,然而在某些情况下它有可能产生“数据饱和”现象.本文提出的U—D固子算法可以在不改变原有算法的基础上克服“数据饱和”现象,同时还能缩短运算时间。

The recursive least squares method is the parameter estimationmethod and it is usuallty applied. However, it can cause the 'data saturation' phe-nomenon. In this peper the U--D factorization is developed for eliminating the 'datasaturation' phenomenon and it not vary the original method. At the same time, it canshorten the operating time.