研究时滞离散系统的线性最小均方差估计.针对具有即时观测和两个延迟观测的线性系统,通过构造重组新息序列,提出了时滞系统的Kalman滤波的一种新算法.其计算归结为三个与原系统有相同维数的标准Kalman滤波器.该方法具有很大的推广应用价值,可用来解决控制理论中一些疑难问题,如H∞固定时滞平滑估计,预演控制及时滞系统的H∞滤波和控制等. To study the linear minimum mean square error estimation of discrete-time systems with time-delay, a new algorithm for Kalman filtering of time-delay systems is proposed for linear systems with real-time observation and two delay observations by constructing a reorganization interest sequence. Three standard Kalman filters with the same dimension as the original system, which have great value in popularization and application and can be used to solve some difficult problems in control theory, such as H∞ fixed delay smoothing estimation, rehearsal control and time-delay systems H∞ filtering and control.