共 32 条
Tikhonov Regularization Based Modeling and Sidereal Filtering Mitigation of GNSS Multipath Errors
被引:22
作者:
Chang, Guobin
[1
,2
]
Chen, Chao
[1
]
Yang, Yuanxi
[2
]
Xu, Tianhe
[2
,3
]
机构:
[1] China Univ Min & Technol, Sch Environm Sci & Spatial Informat, Xuzhou 221116, Jiangsu, Peoples R China
[2] Xian Res Inst Surveying & Mapping, State Key Lab Geoinformat Engn, Xian 710054, Shaanxi, Peoples R China
[3] Shandong Univ, Inst Space Sci, Weihai 264209, Peoples R China
来源:
REMOTE SENSING
|
2018年
/
10卷
/
11期
关键词:
GNSS;
multipath;
sidereal filtering;
Tikhonov regularization;
bootstrap;
GPS;
WAVELET;
EXTRACTION;
D O I:
10.3390/rs10111801
中图分类号:
X [环境科学、安全科学];
学科分类号:
08 ;
0830 ;
摘要:
In Global Navigation Satellite System (GNSS) relative positioning applications, multipath errors are non-negligible. Mitigation of the multipath error is an important task for precise positioning and it is possible due to the repeatability, even without any rigorous mathematical model. Empirical modeling is required for this mitigation. In this work, the multipath error modeling using carrier phase measurement residuals is realized by solving a regularization problem. Two Tikhonov regularization schemes, namely with the first and the second order differences, are considered. For each scheme, efficient numerical algorithms are developed to find the solutions, namely the Thomas algorithm and Cholesky rank-one update algorithm for the first and the second differences, respectively. Regularization parameters or Lagrange multipliers are optimized using the bootstrap method. In experiment, data on the first day are processed to construct a multipath model for each satellite (except the reference one), and then the model is used to correct the measurement on the second day, namely following the sidereal filtering approach. The smoothness of the coordinates calculated using the corrected measurements is improved significantly compared to those using the raw measurement. The efficacy of the proposed method is illustrated by the actual calculation.
引用
收藏
页数:18
相关论文