Automated Generation of OpenSCENARIO Scenarios for Evaluating Advanced Driver Assistance Systems

The automotive industry widely adopts virtual testing of advanced driver assistance systems (ADAS) since it is more reliable, time-saving, and cost-efficient. Constant testing and validation ensures the safety and reliability of a driver-assistance feature. It is initially done on a virtual platform since on-road testing is expensive and cannot reproduce the same results. Testing in a virtual environment uses a scenario-based approach where each scenario represents a real-world condition, and the simulation can help replicate real-world results. Scenarios are usually designed manually by experts following specific rules and standards and is a time-consuming process. The proposed method uses an OpenDRIVE map as its primary input and some user-specified parameters to generate scenarios. Two methods are studied for route generation: one involves route generation using simulator callbacks, and the second method uses an analytical method for calculating the route trajectory for each dynamic entity based on its position in the road network. The output is an OpenSCENARIO file containing information regarding the position, speed, trajectory, and maneuvers of every dynamic entity in the road network. The proposed method gives more control to the user for scenario generation and can automatically generate large scenario datasets quickly compared to manual scenario creation.