SVA: A Street-View-Aided GNSS Positioning Framework With 2DSDM and Likelihood Road for NLOS/Multipath Mitigation

Global Navigation Satellite System (GNSS) suffers severe accuracy degradation in urban environments due to Non-Line-of-Sight (NLOS) and multipath effects. Several methods have been proposed to detect and mitigate NLOS/multipath, but those rely on additional equipment, high costs, and limited multipath detection capabilities. Instead, we propose a framework for a street-view-aided GNSS positioning system (SVA). It fuses 2D shadow matching (2DSDM), likelihood roads, and single-differenced residual line (SDRes Line) via factor graph optimization (FGO) for effective NLOS/multipath detection and mitigation. First, SVA detects NLOS by projecting satellites onto street view images and realizes 2DSDM. Then SVA constrains positioning using likelihood roads extracted from street views. By combining these roads with SDRes Lines, multipath signals are further detected. Finally, FGO integrates all constraints to yield optimized positioning. Experiments on public datasets demonstrate SVA's effectiveness in detecting NLOS/multipath signals. The results show that SVA outperforms the state-of-the-art tightly coupled deep learning/GNSS (TDL-GNSS) and 3D Map Aided (3DMA) methods with maximum improvements of 39.01% in Mean Error (MEAN), 35.99% in Root Mean Square Error (RMSE), 34.47% in Standard Deviation (STD), and 53.34% in Maximum Error (MAX).