摘要
本文通过对工程上常用的残差初值选取方法——均值法统计特性的深入分析,揭示了残差初值问题的重要性.在研究了均值法和最小二乘法等两种简单残差初值估计方法统计特性的基础上,得到了一种简单、实用、附加计算量小、统计特性优良的残差初值估计方法.该方法的特点是能自动根据系统参数的变化来调整残差初值算法的参数,保证残差初值估计的优良统计特性.
The statistics of the average method, a widely used one for estimating initial residuls, is analyzed in detail in this paper. By the statistics, the importance of estimating initial residuls is shown. Based on the comparison between the average and the least square method (the two kinds of simplest methods for estimating initial residuls), a simple applicable method with good statistics and little additional compuration burden for estimating initial residuls is proposed. This method can regulate its parameters to keep the statistics good by itself according to the system parameters.
出处
《自动化学报》
EI
CSCD
北大核心
1991年第4期395-403,共9页
Acta Automatica Sinica
关键词
残差初值
自适应设计
均值法
Initial residuls
residul square sum
adaptive estimation
robust estimation
statistics
