几种最优滤波方法的分析和比较  

Analyses and comparisons of some filter methods

作　　者：黄贤源[1] 隋立芬[1] 范澎湃[1] 

机构地区：[1]信息工程大学测绘学院,河南郑州450052

出　　处：《测绘工程》2007年第3期35-39,共5页Engineering of Surveying and Mapping

摘　　要：解决状态或信号最优估计问题的方法有很多,主要有Wiener滤波、经典Kalman滤波及以白噪声估计理论为基础的现代时间序列分析方法。系统总结了3种主要滤波方法的基本思想、适用条件,并结合算例进行了分析和比较。结果表明,经典Kalman滤波虽然能处理时变系统,但是求解Riccati方程要求较大的计算量,而现代时间序列分析方法则通过构造ARMA信息模型来代替求解Riccati方程,虽减少了计算量,但该方法仅仅适用于处理时不变系统。Many methods to solve the optimal estimation about states or signals have been bought up, including Wiener filter, Kalman filter and modern time series analysis method which is on the ground of white noise estimation principle. It discusses the basic idea and applicability of these methods systematically. Then the analyses and comparisons results are given by an example. It is shown that Kalman filter can deal with time-unsteadily system, but it also brings large calculation in the process of computing the Riccati formula. Modern time series analysis method utilizes ARMA innovation model instead of Riccati formula. It can reduce calculation, but can not treat with time-steadily system.

关 键 词：WIENER滤波 KALMAN滤波 ARMA模型 时间序列分析方法 

分 类 号：P228[天文地球—大地测量学与测量工程]