A generic map based environment representation for driver assistance systems applied to detect convoy tracks

Future Advanced Driver Assistance Systems (ADAS) require generic map based representations to process environment information measured by different sensors. In contrast to today's widespread grid based maps, such environment representations have to be a compact, scalable and easy-to-interpret. In order to fulfill these requirements, we propose a new generic map representation, the two dimensional interval map. One example of information that can be represented in this data structure are convoy tracks. Convoy tracks describe the common motion of vehicles driving in convoys. These tracks can serve as an additional, independent input parameter for longitudinal and lateral control in highly automated ADAS. Experimental results show the general ability of the new data structure to provide sufficiently precise results at low computation time and memory consumption. The developed methodology to detect convoy tracks is capable of extracting multiple common convoy attributes even in complex scenarios with lane change maneuvers.