随机系统最小二乘辨识的稳健性收敛性及适应控制(英文)  

Robustness and Convergence of Least-Squares Identification and Adaptive Tracking for Stochastic System

作　　者：李瑞胜[1] 马平[1] 洪惠民 

机构地区：[1]山东大学威海分校控制工程系,威海264200

出　　处：《控制理论与应用》1992年第3期268-278,共11页Control Theory & Applications

摘　　要：基于估计随机系统已建模部分未知参数的推广最小二乘算法,递推地定义了跟踪随机参考信号的适应控制器。我们证明了闭环系统稳定;当未建模动态特性在平均意义下有界时,估计误差随未建模 动态特性的衰减而减少且跟踪误差以一个微小的量偏离其最小值;当未建模动态特性在平均意义下趋于零时,可同时获得估计的强一致性及适应跟踪的渐近最优性。Based on the extended least squares (ELS) algorithm estimating unknown parameters of modeled part in a stochastic system the adaptive controller for tracking a stochastic reference signal is recursively defined.It is shown that the closed-loop system is stable;the estimation error decreases as the unmodeled dynamics decays and the tracking error differs from the minimum plus a small value when the unmodeled dynamics is bounded in the average sense;the strong consistency of the estimates and the asymptotical optimality of the adaptive tracking are obtained when the unmodeled dynamics approaches zero in the average sense.

关 键 词：随机系统 最小二乘 收敛性 稳健性 

分 类 号：TP271.74[自动化与计算机技术—检测技术与自动化装置]