非持续激励条件下系统辨识递推最小二乘最小范数算法.pdfVIP

控制理论与应用 Vol. 26No. 4 第 26 卷第4 期 2ω9 2佣9 年4 Control 文章编号: 1棚-8152(2(阴阳一0365-06 非持续激励条件下系统辨识递推最小二乘最小范数算法 李银阁，汤卓群，黄锚 (重庆邮电大学自动化学院，重庆4仪灿65) 摘要:系缆辨识中广泛向用的最小二乘算法需要输入向最序列满足持续激励性条件(PE条件); fl3.在大多情况下 这是难以满足的.本文提出了一种不依赖于PE条件的递推簸小二乘、最小范数辨识算法.首先分析了最小二二乘算法 解空间的结构，并i运用罚的数方法，将参数辨识问~转化为无约束优化问舰.然后，提出了将步长、罚因f等过和控 制参数统一的i去代一边推形式的辨识算法，证明了算站在给定的控制参数约束下收敛于唯一的最小λ乘、最小范 数解向最仿真实验表明在非PE条件 F算陆的有效性. 关键词:系统辨识;最小二乘铸怯;持续激胁条件;最小工乘最小姐数解 中图分提号: TP13 文献标识码:A Recursive least-squares and minimum-norm algorithm for system identification without persistent excitation condition LI Yin~guo， TANG Zhuo-qun , HUANG Lei (Department of Automation. Chongqing University of Posts and Telecommunications, Chongqing 4创泊65， China) Abstract: The widely used least附squares algorithms for system identification rely on the assumption that the s叫uenceof input vectors satisfies the persistent excitation conditions (PE conditions); however，伽is conditìon is di俑cu1t to be realized in most cases. For this pu叩ose. a recursive least叫uares and minimum.嗣norm (RLS…MN) identification algorithm that does not depend upon PE conditions is proposed. First, the structure of the solution space of the least 叫uares a1gorithm is analyzed and the parameter identification problem is converted to an unconstraint optimization problem by using the penalty 伽nction method; and 由阳， an iteration咄咄rsion-based identification algorithm is presenωd by unifying the process control parameters , such as st叩 width and penalty factor. This algorithm is proved tωocωonvl刊er嗨ge to a unique 怆1eaωst-翩唱呵squa盯r飞4 and minimum-norm sol