High-performance serving of large-scale OpenDRIVE datasets using standardized GIS technology

Abstract

In addition to prototyping of small-scale road segments for special driving simulation use cases the depiction of real-world road network datasets in OpenDRIVE becomes increasingly prominent. Various German and international test bed activities acquire extensive amounts of highly detailed road network data covering hundreds and thousands of kilometres of motorways, extra-urban and urban roads. Due to OpenDRIVE's complex data structure the plain management of such big datasets is already cumbersome. It becomes even more challenging when - according to changing requirements during project runtime - varying snippets of these datasets are to be extracted dynamically. This subset extraction is a common case because such huge datasets are seldom used in whole for a limited simulation use case. Addressing these challenges, this presentation shows one possible approach for well-performing serving of large-scale OpenDRIVE datasets without any rocket science but with the use of standardized, well-established technologies of the geodata domain. The proposed approach introduces persisting of OpenDRIVE elements in spatial databases through geometry discretisation into OGC Simple Features while maintaining the original XML data for fast subset generation and data browsing. Based on that, open GIS frameworks are presented which enable ad hoc visualization and web-map publishing of such OpenDRIVE data.