利用窗口小波去噪的高精度动态单历元定位算法  

A Precise Kinematic Single Epoch Positioning Algorithm Using Moving Window Wavelet Denoising

作　　者：王德军[1,2,3] 熊永良[1] 徐韶光[1] 

机构地区：[1]西南交通大学测量工程系,四川成都610031 [2]河北工业大学测绘工程系,天津300401 [3]河北省土木工程技术研究中心,天津300401

出　　处：《武汉大学学报（信息科学版）》2015年第6期779-784,852,共7页Geomatics and Information Science of Wuhan University

基　　金：国家自然科学基金资助项目(40874015;41274044;41104007)~~

摘　　要：为提高动态单历元定位的精度和实现实时处理,提出了一种利用窗口小波去噪的高精度动态单历元算法。该算法首先利用常规单历元方法获得观测值残差,在当前历元右侧通过数据延拓来改善端部效应的影响,然后利用窗口小波对当前历元观测值残差去噪,通过交叉证认实现小波去噪分解层数的自适应选取,最后利用去噪后得到的较为干净的观测值实现最终定位。实验结果表明,与常规单历元定位相比,在不同的运动情况下,该方法的定位精度均有较大的提高,与小波后处理结果一致。In order to improve kinematic single epoch positioning accuracy and realize real-time processing, the article presents a precise kinematic single epoch positioning algorithm using moving window wavelet denoising. Firstly, we obtain the observation residuals by using the conventional single epoch method and improve end effect by data extending a part of the current epoch. Secondly, we denoise the current tion of the wavelet epoch observation residuals using a window wavelet, and achieve an adaptive selec decomposition level for denoising through cross-validation. Finally, we achieve final positioning accuracy using the clean observations, after denosing. Comparing the proposed method with conventional single epoch positioning, experimental results show that positioning accuracy has been improved greatly for different motions. The results were consistent with those from wavelet post processing.

关 键 词：单历元 动态定位 多路径 小波去噪 

分 类 号：P228.41[天文地球—大地测量学与测量工程]