Urban multipath detection and mitigation with dynamic 3D maps for reliable land vehicle localization

Abstract

Accurate and reliable positioning is an important prerequisite for numerous vehicular applications. Localization techniques based on satellite navigation systems are nowadays standard and deployed in most commercial vehicles. When such a standalone positioning is used in challenging environments like dense urban areas, the localization performance often dramatically degrades due to blocked and reflected satellites signals. In this paper, a general and lightweight probabilistic positioning algorithm with integrated multipath detection through 3D environmental building models is presented. It will be shown that the proposed system outperforms-in terms of accuracy and integrity-existing methods without introducing additional hardware sensors. Furthermore, a benefit analysis of the suggested 3D model for tightly and loosely coupled GPS/INS sensor integration schemas is provided. Finally, the algorithm will be evaluated with real-world data collected during an urban measurement campaign.