摘要
本文提出了同时估计多变量系统所有参数的递推增广最小二乘法(MRELS),并用Martingale理论分析了算法的收敛性。与文〔1〕的子系统递推增广最小二乘法(SRELS)相比,MRELS的计算量要小得多。数字仿真结果表明MRELS是有效的。
In this paper the recursive extended least squares method for linear multivariable systems is presented and can simultaneously estimate all parameters of the whole system. The convergence of the algorithm is analyzed by means of Martingale theory. Compared with the extended least squares method for subsystems in reference〔1〕, the algorithm proposed in this paper has lessamount of calculation. The simulation results indicate that the algorithm works quite well.
出处
《控制与决策》
EI
CSCD
北大核心
1992年第6期443-447,473,共6页
Control and Decision
基金
国家"八五"项目资金
关键词
多参数系统
最小二乘法
收敛性
multivariable system
recursive identification
Martingale theory