Integration of GNSS positioning and 3D map using particle filter

This paper proposes a novel positioning technique that can be used for vehicles and pedestrians in urban environments in which there are large global navigation satellite system (GNSS) positioning errors because of multipath signals. In GNSS positioning, invisible satellites that are obstructed by buildings emit reflection and diffraction signals, which are called non-line-of-sight (NLOS) multipath signals. These cause major positioning errors in GNSS positioning. Thus, to mitigate the NLOS multipath error, NLOS signals must be identified from all received GNSS signals. 3D environmental maps can be used to identify NLOS signals. In some studies, 3D maps are used to improve GNSS positioning accuracy. However, the previously proposed techniques need a long convergence time to estimate the user position. In this study, we propose a positioning technique that can directly combine GNSS pseudorange-based positioning and 3D environmental maps to improve the positioning accuracy in urban environments where NLOS causes major positioning errors. The key idea behind this paper is to estimate user position using likelihood of position hypotheses computed from GNSS pseudo-ranges which consists of only LOS signals. To determine the NLOS GNSS signals at the user position, it is in turn necessary to accurately predetermine the position before simulation; we solve this problem by using a particle filter. We combine two likelihood estimation methods to determine the likelihood of each particle. The first method uses the Mahalanobis distance and the second method uses the degree of matching of availability of LOS signals. Differing hypotheses of user position are represented by these particles, and for each of these, the likelihood can be evaluated using only the LOS pseudoranges determined from the 3D map. To confirm the effectiveness of the proposed technique, a positioning test was performed in a real-world urban-canyon environment. Using the proposed method, the distribution of the particles converged to within 1 m after 6 s and iterated resampling. In conclusion, our proposed method is suitable and effective for estimating accurate user positioning in urban canyons where conventional GNSS positioning can cause large positioning errors.