Stationary urban environment modeling using multi-layer-grids

Future advanced driver assistant systems and autonomous driving put high demands on the environmental perception especially in urban environments. Today's on-board sensors and on-board algorithms still do not reach a satisfying level of robustness and availability. In this paper we focus on the perception of the stationary environment which mainly consists of raised obstacles and lane-markings. We propose to fuse both the raised obstacles as well as the markings in a grid-based representation for a higher robustness against partial occlusion and false detections. The algorithms are designed to work with both research and close-to-production sensors. The algorithms run in real-time on a standard PC and are evaluated with real sensor data.