Due to the increasing number of research and demonstration projects and the first serial vehicles approaching the market, the safety of automated vehicles (AVs) is widely discussed among many stakeholders, such as academia, governmental organizations, technical services, OEMs, or their suppliers. The fundamental architecture of automated vehicles consists of perception, planning, decision, and actuation. The perception system is responsible for understanding the environment in which the vehicle is inserted and relies mainly on the onboard sensors. Therefore, the available range and vision sensors, e.g., LiDARs, radars, and cameras, have several limitations, such as perception degradation in adverse weather conditions and limited field of view. Questions about how the road infrastructure can collaborate through vehicle-to-x (v2x) communication with the automated vehicles must be clarified to support a high level of acceptance by the individual users on the one hand and the society on the other hand side. This effort should fulfill the vision for mobility in future smart cities. Moreover, comprehending the surroundings is challenging once it is unpredictable and continuously changing. One solution to minimize the onboard sensors' limitations is using infrastructure-based sensors in the form of an intelligent infrastructure unit, also known as a roadside unit (RSU), which can be installed in specific locations and perceive the environment from a different perspective with a higher detection range and field of view. The proposed paper describes the development of an intelligent infrastructure in a real peri-urban road environment, the test field called First Mile Ingolstadt, which focuses on automated vehicles' safety assurance. The Test Field is developed and implemented within the research project IN2Lab. The proposed paper describes the Test Field architecture, the roadside unit concept, the environment perception system, the v2x communication level, and the mission control system. Furthermore, it presents four use cases: traffic monitoring, assisted perception, collaborative perception, and extended perception. The traffic monitoring, based on the perception information provided by each RSU, generates a global fused object list, and monitors the state of the traffic participants. The assisted perception, using vehicle-to-infrastructure communication, broadcasts the state information of the traffic participants to the connected vehicles. The collaborative perception creates a global fused object list with the local detections of connected vehicles and the detections provided by the RSUs, making it available for all connected vehicles. Lastly, the extended environment perception monitors specific locations, recognizes critical scenarios involving vulnerable road users and automated vehicles, and generates a suitable avoidance maneuver to avoid or mitigate the occurrence of collisions. KEYWORDS – automated vehicles, infrastructure-based sensors, test filed, safety.
Mr. Thiago de Borba, MEng., Technische Hochschule Ingolstadt