Automated generation of virtual road scenarios for efficient tests of driver assistance systems

Today, driver assistance systems and autonomous vehicles are actively developed and verified in real-world test scenarios. These real-world tests are neither cheap nor sufficient to ensure a reasonably low enough risk of failure. In order to enable efficient and methodical system tests, we propose the use of Virtual Testbeds (VTBs) as a software environment in which "Digital Twins" of real vehicles and road scenarios are modelled and simulated. In this contribution, we describe a rule-based approach and overall workflow for semi-automatic generation of road scenarios for the automated testing of driver assistance systems. Furthermore, a combination of existing procedural generation techniques suited to this application is selected to automatically generate scenes which adequately mimic real-world scenarios.