Virtual Test Field for Highly Automated Vehicle Systems in Urban Environments

Sammanfattning: Autonomous driving was and is one of the most important research and innovation drivers in the automotive and supplier industry. In addition to the predicted energy savings, a reduction in the number of accidents and their level of damage is also expected. In particular, the functional testing and legislation of highly automated driving functions play a decisive key role here. This results in a justified need for innovation and research and means major challenges for the entire technology sector. Traditional methods such as real-world tests and X-in-the-loop tests for proving functional safety still have their justification, but cannot answer all the questions posed by the diverse requirements in daily use. In particular, urban environments with highly complex traffic scenarios and diverse groups of actors can only be mapped to a limited extent using existing methods.In this work, a novel approach for testing automated vehicle systems in urban environments is presented. The goal is to create a safe and valid environment in which the vehicle under test can interact with real road users under realistic conditions. The basis is a highly realistic virtual model of a German city center. The physical behavior of the vehicle and the pedestrian is measured and transferred to the virtual city model in real time. Sensor models enable the interaction of the vehicle with the virtual environment and the pedestrian. With the help of different studies with different focuses, both individual functionalities as well as the overall functionality are finally evaluated.