As advanced driver assistance systems (ADAS) continuously evolve and develop , their testing standards and descriptions are also changing. The borders between simulation and real-world testing are slowly disappearing. The Budapest University of Technology and Economics, along with the ZalaZONE research group, are working together to develop new, state of the art testing methods, specifically in the space where the virtually created scenarios can be implemented and transferred in real-time into an actual physical scenario test actuating real-world test equipment (like pedestrian target, guided soft target, steering robot) on the ZalaZONE Automotive Proving Ground.
With the development of automotive simulation software, there are more and more possibilities to perform simulations that are very close to reality. However, a simulation is most comparable to reality when the simulation environment has been created with high accuracy. Accurately simulated environments are also referred to as digital twins, as they are complete copies of the real environment. A sufficiently accurate description of the real environment allows for simplifying the validation flow of the simulation results.
ZalaZONE Proving Ground shares the HD map of its different modules to support this process. These HD maps contain all the information that current technology conditions allow, such as the exact geometric positioning of roads and their surroundings, the longitudinal and transversal alignment of roads, and the highly accurate positioning of markers. Maps produced from LiDAR point clouds have an accuracy of 2 cm, and the coordinates recorded during the survey are accurate in the world coordinate system. The available maps are of great value to testers and developers, as there are currently few maps of this accuracy available completely free of charge to anyone. The maps have been created to the increasingly popular OpenDRIVE standard, ensuring compatibility with current and future software. With these maps, we give testers the opportunity to test their camera and LiDAR systems, for which the HD map can serve as ground truth, and to create different scenarios in their own simulation environment based on our HD maps, which they can then execute on the test track with the help of our team.
Several software packages can already provide scenario generation using OpenDRIVE files, but VTD Vires provides the most comprehensive solution. When using the VTD Vires scenario editor, importing an OpenDRIVE file will display a bird’s-eye view of the selected map. On this view a selected Euro NCAP “cut-out scenario” can be created. Building a scenario is a relatively quick and easy process: routes are created with spline, clothoid, or polyline lines. The actions are responsible for the test scheme’s events, like lane change or desired target speed of the vehicle. After building the scenario, it can be played in a 3D view and then have the scenario file exported. This results in creating OpenSCENARIO, the simulated scenario pair of the OpenDRIVE road network description, which defines the dynamic elements and respective actions placed on the road network.
In most cases, the scenarios created include participants other than the test vehicle, which are easy to create in the simulation environment, but more difficult to accurately implement in the real world. In a well-defined scenario, it is necessary that all participants can be properly controlled remotely, and that a safe outcome is guaranteed for all test participants, without any injuries or material damage. ZalaZONE Proving Ground can provide the right technological background for such tests with AB Dynamics tools.
Our development team has developed a solution to convert the available OpenDRIVE and OpenSCENARIO files into a format readable by AB Dynamics test tools using Matlab software. Based on the available GPS coordinates, a local origin is calculated, relative to which the positions of each participant are given. Based on the speed profile for each participant, a time-dependent distance traveled curve is generated, and a path curve is available, consisting of the XY coordinates and the path curves. The software created ensures that tests with one or more participants can be handled properly so that more complex scenarios are not a problem. The files created can now be easily imported into AB Dynamics tools software to run the scenario.
As mentioned, the PMC files can be read and even be edited by the AB Dynamics software. Here the speed profile can be checked, and minor errors can be corrected manually. The center of the coordinate system is calculated during the conversion, specified together with the underlay and positioned according to GPS standards and usage. This way, all the routes and their respective positions belonging to the scenario are loadable into the common coordinate system. Also using these routes in simulation, a preliminary offline Collision Check can be run, so as to see the timely course of the scenario’s participative elements on the simulated proving ground track in a simple visualized manner. If this Collision Check result is the same or similar as the simulation created earlier, the development team then can perform the test with real test tools in the physical environment. Furthermore, this scenario coordinate data can also be loaded for vehicles with steering and pedal robots and platforms carrying dummies and therefore allowing an even more complex scenario to be implemented. The whole process of pre-building the scenario in a simulation environment and converting it with a few clicks using a Converter greatly speeds up the compilation of real tests. The design phase can take place immediately in the simulation phase and it is not necessary to structure the route separately in each scenario section for each test participant regarding AB Dynamics software. Lastly, another benefit of using our converter is that there are no pre-measurements to set the scenario’s coordinate system, all of which can be specified in advance where new files can be generated quickly even after multiple modifications.
Written by Testing and Validation Team.