Highly Accurate Scenario and Reference Data for Automated Driving

Authors

L. Vater, R. Krajewski, L. Eckstein, ika, RWTH Aachen University; T. Moers, J. Bock, fka GmbH

Summary

Highly automated driving functions are increasingly based on data-driven approaches. Reference data for environment perception sensors of automated vehicles and trajectory data of real traffic is thus crucial for both the development and the validation of the systems. Within the development process, these are detections of other road users and the environment around the test vehicle, which provides an accurate reference for the onboard sensor technology. For the validation process, naturalistic scenario data is required, which can then be used to identify relevant scenarios for testing the developed system. Measurement methods based on camera-equipped drones are suitable for both applications. It is characterized on the one hand by an advantageous perspective and on the other hand by its mobility compared to other measurement methods.