Advanced preprocessing of GNSS data for traffic analysis using the hampel filter, hybrid interpolation models, and the kalman filter
Transport and Communications Science Journal · 04.01.2026 · 77 (3) · 328-341
DOI: 10.47869/tcsj.77.3.8 ↗Kurz gefasst
Fachartikel in der Zeitschrift Transport and Communications Science Journal, online erschienen am 04.01.2026. Verfasst von Mạnh Đỗ Văn.
Abstract (Originalsprache)
GNSS-mounted data collected from vehicles has become a critical source for urban traffic research and management due to its ability to continuously capture position, speed, and trajectory in real time, thereby supporting the estimation of traffic density, average speed, travel time, congestion detection, and mobility behavior analysis. However, in urban environments, GNSS accuracy is often degraded by multipath effects, satellite signal obstruction caused by high-rise buildings, and device limitations, leading to noisy speed measurements, outliers, and missing data, which can significantly bias results if raw data are used directly. This study proposes a three-stage preprocessing framework: (i) applying a Hampel filter to detect and replace outliers based on the median; (ii) reconstructing missing data using conditional linear interpolation for short gaps, combined with neighboring value propagation methods (LOCF and NOCB) for boundary segments; and (iii) employing a Kalman filter to smooth the speed time series and eliminate near-zero values caused by GNSS noise. The results demonstrate that the proposed approach effectively restores a continuous speed profile, reduces measurement noise, and enhances data reliability for subsequent urban traffic analysis.
Zitieren
APA
Văn, M. Đ. (2026). Advanced preprocessing of GNSS data for traffic analysis using the hampel filter, hybrid interpolation models, and the kalman filter. Transport and Communications Science Journal, 77(3), 328-341. https://doi.org/10.47869/tcsj.77.3.8
Vancouver
Văn MĐ. Advanced preprocessing of GNSS data for traffic analysis using the hampel filter, hybrid interpolation models, and the kalman filter. Transport and Communications Science Journal. 2026;77(3):328-341. doi:10.47869/tcsj.77.3.8
BibTeX
@article{Vn2026,
author = {Văn, Mạnh Đỗ},
title = {Advanced preprocessing of GNSS data for traffic analysis using the hampel filter, hybrid interpolation models, and the kalman filter},
journal = {Transport and Communications Science Journal},
year = {2026},
volume = {77},
number = {3},
pages = {328--341},
doi = {10.47869/tcsj.77.3.8}
}
EUVIMED