Infrastructure-aided Automated Driving in Highly Dynamic Urban Environments

Abstract

Automated driving without driver involvement requires that the automated driving system (ADS) is able to detect relevant environmental factors at any time and to make a prediction of the situational evolution within the planning horizon on this basis. However, in unclear traffic areas, a comprehensive detection from the first-person perspective cannot be guaranteed, which may result in safety issues, a defensive planning behavior of the automated vehicle and thus reduce the traffic flow. A system is presented that generates collision-free paths based on single-object predictions and a cyclic trajectory planning process, incorporating a traffic infrastructure, enhanced with sensors, processing and communication units. Within this work, the usefulness of infrastructure extensions for the support of connected and automated vehicle (CAV) is investigated in a complex intersection scenario with the help of the above-mentioned methods.