多级分解预测与滤波方法

本文在Andrisani等人近年来工作的基础下,进一步发展了线性离散随机系统的多级分解估计方法,提出了比较完善的多级分解预测与滤波公式,并将本文的工作与Andrisani等人的工作作了比较。本文给出的预测与滤波算法计算量较小,便于实时处理,且估算精度与集中估计方法基本一致。文中给出了应用算例,分析了模型误差对预测和滤波结果的影响。

Multistage Prediction and Filtering Methods Using Partitioning

In this paper, a relatively prefect prediction a and filtering algorithm is obtained, based on the recent work of Andrisani, for linear discrete time stochastic system. A comparison is made of the work in this paper with Andrisani's work. The results of Andrisani's could be considered as a special case of the work in this paper. The algorithm consists of two consecutive estimators. Further discussion is made of the interconnection between this structure and the more familiar one stage optimal Kalman estimator.In the end a numerical example is given for illustrating the results of this paper.