ForschungPublikationen
Detail-Ansicht

Quantification of GNSS NLOS Spatial Correlation

A Case Study in Hong Kong's Urban Canyon

verfasst von
Lucy Icking, Guohao Zhang, Li Ta Hsu, Steffen Schön
Abstract

In urban environments, non-line-of-sight (NLOS) signal conditions are a major error contributor for GNSS positioning, the signal delay can reach up to more than one hundred meters due to reflection in urban canyons such as Hong Kong. Collaborative positioning between traffic participants can help to eliminate common errors, but comprehensive analyses on how to find common errors in cities and how to quantify them are required. In this study, we present a new method for finding locations in urban areas with similar extra path delays on the basis of ray-tracing with 3D building models, exemplarily for an urban canyon in Hong Kong. Using the 2D Pearson correlation coefficient, we calculate the spatial correlation of reflection extra paths at two locations to generate a similarity measure. With this measure, we quantify the amount of similar errors at two locations. In realistic simulations, we show that two locations with highly correlated extra path delays show better results concerning single-difference error correction and relative positioning errors. The single differences show that for the selected area in Hong Kong, the higher the spatial correlation, the higher the amount of common extra path delays. Furthermore, we are able to show that the mean relative positioning error can be reduced from 42.4 m for a low correlation pair of agents to 12.7 m for a high spatial correlation pair.

Organisationseinheit(en)
Institut für Erdmessung
Externe Organisation(en)
Hong Kong Polytechnic University
Typ
Aufsatz in Konferenzband
Seiten
712-722
Anzahl der Seiten
11
Publikationsdatum
2022
Publikationsstatus
Veröffentlicht
Peer-reviewed
Ja
ASJC Scopus Sachgebiete
Luft- und Raumfahrttechnik, Elektrotechnik und Elektronik
Elektronische Version(en)
https://doi.org/10.33012/2022.18165 (Zugang: Geschlossen)
 

Details im Forschungsportal „Research@Leibniz University“