摘要
本文讨论了参数估计和输出误差的收敛性,并给出了系统持续激励的一个充分条件。在此条件下,证明了参数估计和它与时变参数间的估计误差的均方值收敛于随机参数的期望值和均方差值。
This paper presents new development of parameter identification for MIMO discrete stochastic linear systems with random parameters based on the least squares method. Convergence of both parameter estimates and output error are studie'd and preliminary persistent exciting conditions are given. Under some conditions, it is proved that parameter estimates and mean square value of difference between parameter estimates and these random parameters are ensured converging to expectations and the mean square derivation value of random parameters, resectively.
关键词
系统辨识
线性系统
收敛性分析
system identification
linear systems
convergence analysis
stochastic systems
time-varyingsystems
