Abstract
Recently, autonomous driving vehicles, such as those developed for the DARPA Urban Challenge, have been researched all over the world. In Japan, some autonomous vehicles such as those developed for Tokyo Motor Show 2011 held at last year, have also been developed. For such autonomous vehicles, perception of surroundings around the vehicle is one of the most key technical issues. In the field of perception for the autonomous vehicle, it is important to detect static obstacles and reveal drivable region. Moreover, it is also necessary to detect dynamic objects and to predict its future trajectory, since traffic environment is highly dynamic environment. In this paper, we propose a method to percept surroundings using LIDAR. In this method, obstacles and drivable area is recognized base on Occupancy Grid Maps, which is the method to estimate posterior probability of being occupy. Moreover, dynamic objects are detected based on the Occupancy Grid Maps.
