Zhong, Qiming; (2024) Investigation of 3D-Mapping-Aided GNSS Navigation in Urban Canyons. Doctoral thesis (Ph.D), UCL (University College London).

Preview

Text qm_thesis.pdf - Accepted Version Download (17MB) | Preview

Abstract

In urban environments, the propagation of satellite signals is often affected by buildings, leading to challenges such as multipath interference and non-line-of-sight (NLOS) reception. These challenges compromise the performance of conventional GNSS positioning in cities. Several studies have shown that 3D mapping data significantly improves GNSS positioning by predicting which signals are line-of-sight (LOS) and which are NLOS. This research builds on a new implementation of UCL's 3D-mapping-aided (3DMA) GNSS algorithms in MATLAB, encompassing satellite visibility prediction, shadow matching, likelihood-based ranging, and solution integration, demonstrating improved GNSS positioning performance in cities. Building on these core algorithms, this research introduces several enhancements, including advanced satellite visibility prediction for overhanging structures, the incorporation of untracked satellites in shadow matching, the integration of Bayesian inference with shadow matching to adapt to varying urban densities, a new NLOS model tailored for likelihood-based ranging, and a clustering algorithm based on region growing to handle ambiguity. These enhancements collectively improve 3DMA GNSS performance by about 20%. Additionally, an outlier detection method mitigates the effects of mapping inaccuracies, reducing the positioning error by approximately 15% in both single- and multi-epoch cases, compared to methods without such detection. 3DMA GNSS is combined with particle and grid filters to enable multi-epoch positioning. These 3DMA GNSS filters are most beneficial for mobile positioning in dense urban areas, exhibiting a performance improvement of about 55% over conventional continuous positioning algorithms. This research also introduces an efficient NLOS mitigation algorithm that empowers conventional GNSS algorithms to handle asymmetric distributions of NLOS measurement errors, without requiring supplementary information or hardware. This algorithm demonstrates improvements of 12% and 34% in single- and multi-epoch cases, respectively, compared to conventional positioning methods. In summary, this research offers a comprehensive suite of advancements in 3DMA GNSS positioning, particularly in challenging urban landscapes, with promising implications for real-world applications.

Download activity - last month

Export as JSONCSVXML

Download activity - last 12 months

Export as XMLJSONCSV

Downloads by country - last 12 months

Export as XMLJSONCSV

Archive Staff Only

View Item