摘要
本文提出了一种新的具有递推性的结构辨识方法.该方法使用系统正常运行的测量数据,在载噪且噪声统计特性未知及数据较少时,可以准确地确定系统的结构不变量,同时得到系统参数的最小二乘估计值.数值实例中的仿真计算表明该方法是可靠的.
In this paper, a new recursive structural identification method for the discrete linear multivariable systems is proposed. By using the data of the systems in the normal operation along with some unknown statistic properties, this method can exactly determine the structural invariants of the systems. The data required is much less than those in [1—4, 6—14]. At the same time, the least square estimates of the system parameters are also obtained. The results of the simulation example show that this method is reliable.
出处
《自动化学报》
EI
CSCD
北大核心
1991年第5期571-576,共6页
Acta Automatica Sinica
关键词
多变量系统
结构辨识
线性系统
Structural identification
canonical form
invariant
