Collaborative Localization with Truth Discovery for Heterogeneous and Dynamic Vehicular Networks

Collaborative localization over vehicular networks is challenging if quality varies among the collected multiple sources of information. These information sources are either from vehicle on-board sensors or remote sensing using vehicular communications. The variation in the quality of remote sensor information may cause estimation performance deterioration, even threatening the system security. In this paper, we propose a distributed localization framework with truth discovery for heterogeneous and dynamic vehicular networks. Firstly, it allows vehicles to learn which neighboring vehicles they should cooperate with. Secondly, it is resilient against the quality variation of the shared information between the connected vehicles.