Improving Road Network Surveying in the Automotive Development Process through Road2Simulation Guidelines

Abstract

The test and development of driving assistance and automation systems needs map data that is more detailed and includes more additional information than needed later on in the driving function. Gathering real-world data with all necessary details is still a job for professionals and thus expensive and time consuming. On the contrary, public authorities are creating and updating enormous amount of data to manage their cities and road infrastructure. It would be very helpful if such public data - which is collected anyway - was facilitated for the automotive development process and if generation of such data was more time and cost efficient. To reach this goal German OEM Audi, BMW, Daimler, Porsche and Volkswagen collaborated with the German Aerospace Center (DLR) to create guidelines for surveying roads and a lightweight data model to store this information supporting both domains: automotive test and development as well as city administration. The guidelines tell data processors on the one hand how to handle various road situations such as motorways, complex intersections, roundabouts etc. and on the other hand how road infrastructure such as traffic islands, signs and road furniture should be handled. An important part is the attribution of the geo-objects with metadata for an automated transformation into complex road description formats (such as OpenDRIVE, or Navigation Data Standard). The data model is designed as intermediate data format and for easy management and processability with common GIS tools (e.g. QGIS and ArcGIS in combination with PostgreSQL or Oracle Spatial). The guidelines and the data model are made available for free and are further maintained by DLR. Road2Simulation was already used to create a test course in Stuttgart downtown and is a basis for the digital representation of German test field activities in Düsseldorf (project "KoMoD") and Braunschweig with its environs (projects "AIM" and "Testfeld Niedersachsen") as well as in Graz and its environs (project "ALP.Lab") in Austria. Therefore Road2Simulation was recently updated to version 1.2 with improvements in the data model and a detailed handling of road infrastructure. Road2Simulation helps to reduces costs because the surveyor does not have to become familiar with different complex road description formats. He has only to digitalize the road data following the guidelines, which is usually part of his standard business. The data acquirer is more flexible in using the data regarding a target format and accuracy of the modeling. If public authorities provide their data in Road2Simulation format, more customers can make use of it and more complex real-world data will be available for test and development of automated driving functions.