面向复杂条件的北斗/GNSS数据粗差处理方法  

BDS/GNSS data outlier processing approach for complex conditions

作　　者：李雪珍 章浙涛 刘欢 袁海军 LI Xuezhen;ZHANG Zhetao;LIU Huan;YUAN Haijun(School of Earth Sciences and Engineering,Hohai University,Nanjing 211100;Shanghai High Gain Information Technology Co.,Ltd,Shanghai 201700)

机构地区：[1]河海大学地球科学与工程学院,南京211100 [2]上海海积信息科技股份有限公司,上海201700

出　　处：《南京信息工程大学学报（自然科学版）》2022年第6期722-730,共9页Journal of Nanjing University of Information Science & Technology（Natural Science Edition）

基　　金：国家自然科学基金(42004014);江苏省自然科学基金青年基金(BK20200530);中国博士后科学基金(2020M671324);江苏省博士后科研资助计划(2020Z412)。

摘　　要：所有观测值中不可避免含有粗差,在具有高遮挡和强反射特性的峡谷环境以及多频多系统的复杂条件下,粗差问题尤为突出.因此研究面向复杂条件时的北斗/GNSS数据粗差的探测识别,以确保定位结果准确可靠.基于均值漂移和方差膨胀的粗差探测提出不同阈值的方案,利用实测数据进行粗差探测实验,分析面向复杂条件时的各方案粗差处理的效果.实验结果表明,基于均值漂移的小阈值(U_(1/2α2)(0,1))方案比无粗差处理方案的定位误差RMSE在E、N、U方向上分别改善了0.059 m、0.017 m和0.062 m.基于方差膨胀的大迭代阈值(ω)结合小阈值(k_(0)、k_(1))方案比无粗差处理方案的定位误差RMSE在E、N、U方向上分别改善了0.098 m、0.055 m和0.209 m.结果表明,在复杂条件下基于均值漂移的小阈值(U_(12α2)(0,1))方案和基于方差膨胀的大迭代阈值(ω)结合小阈值(k_(0)、k_(1))方案能够更好地探测和识别粗差.All observations inevitably contain outliers.Outlier detection is a particularly tough problem under the complex conditions of the combination of multi-frequency and multi-constellation and urban canyon environments.Here,outlier detection is studied to ensure accurate and reliable positioning performance of BDS/GNSS data.The outlier detection schemes with different thresholds were proposed based on mean shift and variance inflation.Then dedicated experiments under complex conditions were carried out to evaluate the performance of the proposed outlier detection schemes.The experimental results show that the RMSE of the mean shift based small threshold(U_(12α2)(0,1))scheme is improved by 0.059 m,0.017 m,and 0.062 m in the E,N,and U directions,respectively.The RMSE of the combined large iterative threshold(ω)and small threshold(k _(0)、k _(1))scheme based on variance inflation is improved by 0.098 m,0.055 m,and 0.209 m in the E,N,and U directions,respectively.The results show that the above two schemes can detect and identify outliers under complex conditions.

关 键 词：粗差处理 复杂条件 稳健估计 数据探测 

分 类 号：P228[天文地球—大地测量学与测量工程]