Improving accuracy and integrity with a probabilistic Urban Trench modeling
Résumé
GNSS localization in urban environment remains a key issue that must be solved in order to enable innovative driver assistance systems or cooperative applications. This positioning key issue exists not only in terms of accuracy, but also in terms of integrity. Actually, one aims at reducing as much as possible the error mainly due to multipath on buildings, and provide in the mean time a certain guarantee on the user position exactness. This communication shows how a simple geometrical modeling, named 'Urban Trench', of the city environment can improve significantly the positioning accuracy. Authors also propose an algorithm for computing in this context an urban trench protection level, with a much higher reliability compared to standard protection level whose open sky assumption is violated. Preliminary results are given in the city of Nantes, France. One of the most interesting feature of the model is that it complies with the existing digital road map features and does not cause data storage or computational throughput problems.