Most of the Intelligent Transport Systems rely on GNSS positioning information. Unfortunately, most of their uses occur in dense urban areas where signal propagation is degraded by surrounding obstacles because of the reception of NLOS (Non Line-Of-Sight) signals. Since the end of the 90’s, as a research institute concerned by transport applications, the LEOST team of IFSTTAR pursues research on the impact of the surroundings of a GNSS antenna on the receiver’s performances and in particular on NLOS cases when the direct path is not present or it is undetectable due to its low power compared to the rest of path. NLOS detection and mitigation remains an open research problem that can be studied from different angles: at signal processing level, at observable level and at PVT level. First works of LEOST have dealt to the development of a satellite availability prediction tool based on image processing already (PREDISSAT) for guided transport modes. Then, in collaboration with the Ecole Centrale of Lille, signal processing techniques have been developed to model the evolution of the satellite states of reception versus time and propose pseudo-range error models linked to the different states and in particular the NLOS case where reflected signal can be received without a direct ray. Recent activities are performed in the context of the CAPLOC research project. The main objective of CAPLOC, in collaboration with the Technical University of Belfort Montbéliard, is to provide an innovative tool for the positioning function, relying on satellite-based technologies, GPS and EGNOS, and mitigating the difficulties linked to the constricted environment of reception. In the same way than users of 3D models, our goal is to compare satellite positions to obstacles in order to detect satellite states of reception. Our approach relies on video records of the environments. This paper presents how the images sequences provided by several cameras installed on the roof of the vehicle allow characterizing GNSS performance and gives a quick overview of how they are used for accuracy enhancement in the CAPLOC project.
