摘要
系统地讨论了SISO、线性时不变、指数稳定系统在最坏情况下的l1鲁棒辨识问题。提出了系统模型集合的最小外框概念;建立了两种任意非零信号作用下l1鲁棒辨识算法;提出了任意非零信号作用下系统的可辨识条件;证明了算法的全局收敛性和最优性。
The l1robust idenfification for SISO linear shift invarant exponentially stable discrete systems in worst case is studied in this paper. Present concept of minimal outer box for system models set; Construct two algorithms for l1 robust identification as employed by arbitrarily nonzero signal; Present conditions of identification ability as employed by arbitrarily nonzero signal; and the global convergence and optimality of algorithms are proved.
出处
《控制与决策》
EI
CSCD
北大核心
1995年第3期204-209,共6页
Control and Decision
基金
国家自然科学基金
湖北省自然科学基金
关键词
鲁棒辨识
线性系统
鲁棒控制
系统辨识
l1 robust identification, worst case, central algorithm, interpolatory algorithm, global convergence