GNSS Outlier Mitigation via Graduated Non-Convexity Factor Graph Optimization

Accurate and globally referenced global navigation satellite system (GNSS) based vehicular positioning can be achieved in outlier-free open areas. However, the performance of GNSS significantly degraded by outlier measurements, such as multipath effects non-line-of-sight (NLOS) receptions arising from signal reflections buildings. Inspired advantage batch historical data resisting this paper, we propose a graduated non-convexity factor graph optimization (FGO-GNC) to improve performance, where impact outliers is mitigated estimating optimal weightings measurements. Different existing local solutions, proposed FGO-GNC employs non-convex Geman McClure (GM) function estimate measurements via coarse-to-fine relaxation. The effectiveness method verified through several challenging datasets collected urban canyons Hong Kong using automobile level low-cost smartphone receivers.