Concept of Infrastructure Based Environment Perception for IN2Lab test field for Automated Driving

Abstract

Advanced sensors like Radar, LiDAR and camera are used in development for autonomous vehicles for environment perception to do real time object detection, classification and tracking of dynamic objects. The process of environment perception is a crucial step in development of autonomous vehicles but it is unfortunately, one of the most challenging processes because the outside environment is unpredictable, continuously changing and full of uncertainties. In this situation, the maximum assistance to such vehicles is of utmost need. Development of intelligent infrastructure unit, known as road ride unit (RSU), which can be mounted at various specific locations on the road sides to perceive environment in real time from high angle view can assist automated driving. Such RSUs can consist of various types and combination of sensors and vehicle to infrastructure (V21) communication unit. It can send the information about critical situations, traffic alerts and other important information to nearby vehicles in real time to further enhance automated driving. This paper describes the details of development of intelligent infrastructure for a test field within the research project IN2Lab. test field development. It describes the architecture of RSUs and the concept of environment perception including sensor requirement and selection, sensor field of view, sensor coordinate system and concept of sensor fusion.