摘要
本文扩展了用于分析时不变系统辨识算法收敛性的鞅收敛定理（MCT），建立了鞅（martingale）超收敛定理（MHCT）．它可以作为工具来分析时变系统的各种辨识算法的收敛性，为解决时变系统收敛性和稳定性分析这一困难课题提供了新方法，开辟了新路．本文以遗忘因子最小二乘算法为例，成功地用MHCT分析了它的参数估计的收敛性．
In this paper,the martingale convergence theorem used to analyze the convergence ofidentification algorithms of time-invariant systems is extended. The martingale hyperconvergence theorem(MHCT)is established,which may analyze the convergence of various algorithms for time-varying systems and give anew method for analysis of covengence and stability of time-varying systems. Taking the forgetting factorleast squares algorithm (FFLS) as an example,we prove the convergence of the FFLS algorithm by means ofMHCT.
出处
《控制理论与应用》
EI
CAS
CSCD
北大核心
1997年第1期90-95,共6页
Control Theory & Applications
基金
863应用工程资助
关键词
鞅收敛定理
参数估计
遗忘因子
最小二乘法
time-varying system
martingale convergence theorem
martingale hyperconvergence theorem
parameter estimation
forgetting factor least squares algorithm
