Strategien zur selektion von satelliten in kinematischen gnss-anwendungen auf basis von 3d-umgebungsmodellen

In the area of GNSS positioning, site-dependent errors, which are caused by the antenna environment, are one of the accuracy limiting factors. In this paper a strategy for satellite selection is presented, which considers site-dependent effects like signal diffraction, Non-line-of-sight (non-line-of-sight, NLOS) reception and far-field multipath. The geometry-based determination and the subsequent exclusion of influenced signals is performed using georeferenced 3D point clouds and LoD2 city models. First of all, the application to a static GNSS measurement shows the general potential of positioning with satellite selec-tion, which was particularly noticeable by an increase of the fixed carrier phase ambiguities. This is fol-lowed by the kinematic GNSS application with an additional comparison between the environment models from a TLS survey and a LoD2 city model. In a direct comparison of the impact on GNSS positioning, both approaches achieve the almost identical result, so that none of the methods can be considered to be significantly better. Therefore, the essential difference is the data acquisition, which for georeferenced TLS point clouds of larger measurement areas is associated with a comparatively high effort. The application of both approaches to GNSS positioning offers a significant improvement in trajectory estimation, which can be quantified with an RMS of less than 5 cm compared to the reference trajectory. Furthermore, an additional selection of signals influenced by multipath effects was performed using Fresnel zones, which allows a further improvement of the estimated trajectory.