摘要
迄今为止的辨识或参数估计算法,在系统受到非平稳噪声干扰时往往缺乏一致性,甚至不收敛。本文采用结构补偿技术,提出了系统辨识的模型扩展方法。针对线性系统在较为广泛的一类非平稳干扰下的参数辨识问题,导出了几种适用的递推估计算法,并讨论了收敛性。仿真结果表明,这些算法具有明显的优良品质。
The existing system identification or parametor estimation algo-rithms always suffer from bad consistence and convergence when the system is disturbed by non-stationary noises.In this paper, the structure compensation technique is used, and the model extension methods are developed. Several recursive algorithms for parametor identification of linear system in general non-stationary disturbances are derived, and the convergence of the algo-rithms is discussed. Simulation results showed that the algorithms possess obvious good property.
出处
《控制与决策》
EI
CSCD
北大核心
1990年第2期7-12,24,共7页
Control and Decision
