Improved GNSS Positioning in Urban Environments Using a Logistic Error Model

Zhengdao Li, Penggao Yan, Li-Ta Hsu

Peer Reviewed

Abstract:	A Gaussian error assumption is predominantly utilized in the pseudorange measurement model for global navigation satellite system (GNSS) positioning. However, this assumption is frequently violated in urban environments due to multipath effects and non-line-of-sight (NLOS) receptions, which induce heavy-tailed error distributions. In this paper, we propose to characterize the urban GNSS pseudorange errors using a logistic distribution, which better fits the empirical errors while maintaining a simple parameterization. Based on the logistic error model, we derive a maximum likelihood estimator (MLEr), termed the least quasi-log-cosh (LQLC) estimator. We also propose to solve the LQLC estimation efficiently through an iteratively reweighted least squares (IRLS) solver. Experimental results in light, medium, and deep urban environments demonstrate that the proposed LQLC estimator significantly outperforms the conventional Gaussian-based least squares (LS) estimator, reducing the 3D root mean square error (RMSE) by approximately 11%-31% and the 3D error standard deviation (STD) by up to 60%. The proposed method also maintains high computational efficiency suitable for real-time navigation applications.
Published in:	Proceedings of the ION 2026 Pacific PNT Meeting April 13 - 16, 2026 Hilton Waikiki Beach Honolulu, Hawaii
Pages:	379 - 388
Cite this article:	Li, Zhengdao, Yan, Penggao, Hsu, Li-Ta, "Improved GNSS Positioning in Urban Environments Using a Logistic Error Model," Proceedings of the ION 2026 Pacific PNT Meeting, Honolulu, Hawaii, April 2026, pp. 379-388. https://doi.org/10.33012/2026.20582
Full Paper:	ION Members/Non-Members: 1 Download Credit Sign In