Data Analysis for Automated Driving
The analysis toolbox will help Volkswagen Commercial Vehicles provide data for their Automated Driving Systems.
The Operational Design Domain describes the constraints under which a driving automation system is designed to operate. This includes environmental and geographic information as well as the required presence or absence of certain traffic or roadway characteristics. At Volkswagen Commercial Vehicles, a dedicated team defines such boundary conditions for the development and testing process. This allows the team to evaluate potential service areas for Mobility as a Service applications. Based on our expertise in geospatial data analysis and efficient processing of large volumes of geodata, Volkswagen Commercial Vehicles contracted 52°North to develop a toolbox that enables the automatic derivation of these characteristics from available spatial datasets.
The content-rich road network data provided by Open Street Map builds the foundation for the analysis toolbox. It enables the derivation of basic properties such as the distribution of road types and street fruniture, speed limits or number of lanes for a given area of interest. The toolbox will also support complex analysis capabilities, such as statistics on curve radius, the intersection impact angle, or intersection complexity. The system combines two different analysis concepts:
- Basic analysis using the Overpass Turbo API
- Complex statistics using the Python library osmnx in combination with a PostGIS database
During 2023, the team focused on more complex analysis. Important parameters such as the visibility range of certain roads, the incident angle of roads meeting at an intersection, or the minimum curve radius were derived. Our team integrated all analysis features into one executable toolbox. It can be used from the command line, but is already prepared to run as a service in a larger deployment environment.
Customer
Volkswagen Commercial Vehicles, Germany