Vehicle Localization Using Infrastructure Sensing

Infrastructure supported autonomous driving is getting in the focus of current research. In this work, we investigated the usage of a traffic monitoring infrastructure combined with the environmental model of an autonomous driving vehicle for localization. The forwarded environmental model of the infrastructure contains tracked road users, but the used tracking algorithm is unknown. The result is based on a two-stage transformation process with optimized fusion and tracking by a Kalman filter. Experiments show, that the algorithm provides a consistent localization at the first time step.