Fisheye camera-based local effects mitigation for better accuracy in land transportation applications

GNSS is now deployed in every vehicle but despite all efforts to improve performance, local effects that degrade satellite signal reception still remain a challenge for land transport applications, especially when accuracy is required. State of the art solutions proposed different weighting schemes for enhanced WLS accuracy based on C/N0 and Elevation, alone or combined. We proposed in [3] the added use of a fish-eye camera for LOS/NLOS detection and their associated weights. In this paper, we try to fill in some of the gaps in our method, and in particular to consider an additional case of reception: reception through vegetation, thanks to a news segmentation of our images. A methodology to estimate specific weights for each of these reception states is proposed, based on different pseudorange error models, that are applied in WLS to be compared to state of the art. The results show that the support of the camera can enhance accuracy in specific types of environments (urban or vegetation) but that one global weighting scheme does not satisfy all of them.