Urban automated driving
Due to more complex intersections and road geometries as well as frequent interactions with pedestrians and non-motorized vehicles, urban environments are particularly challenging for automated driving. The members of the @CITY research project attempt to tackle these challenges and test prototype cars under realistic conditions.
Automated vehicles benefit from the availablility of highly accurate map data, as was previously demonstrated in Ko-HAF. As a subcontractor of several members of the @CITY consortium, FORWISS provides tools for parsing and processing HD maps in the OpenDRIVE standard format. In particular, we produce compact, tileable map datasets based on Protocol Buffers.
Within these datasets, linear elements are represented as arc splines – a conceptually simple type of spline which offers performant, closed-form solutions for many commonly required calculations. We use SMAP to produce arc splines, an algorithm which achieves a target accuracy with the minimum possible number of spline segments.
More information on @CITY is available from the official project website.