A Generic Validation Scheme for real-time capable Automotive Radar Sensor Models integrated into an Autonomous Driving Simulator

Since Advanced Driver Assistance Systems are getting more and more complex a strong effort is put into the development of open-source autonomous driving simulators. However a robust virtual environment should not only be assessed by its realistic physics or 3D assets variety, but it also has to incorporate highly accurate and real-time capable sensor models. The research presented in this paper introduces a robust method for validating radar sensor models enabling a simple proof of their reliability in a virtual scenario. To give an overview of radar sensor modeling approaches, the solutions available in the literature are evaluated in terms of usage in virtual environments based on simple criteria. Finally, an exemplary radar sensor model integrated into the CARLA simulator is presented and also a short outline of further research is depicted.