Evidenzorientierte Ableitung von Grundszenarien für die Bundesautobahn

Goal of this research project is the development of a scenario concept aimed at a complete description of the driving task on a German highway. Taking into account existing approaches, a structure based on three pillars is derived. Stationary driving states form the first pillar, temporally extended transitions between these states form the second, and superimposed interaction provide a further way to increase the descriptiveness. Out of these granular building blocks, more complex traffic scenarios can be composed. Safety-relevant scenarios are defined from base scenarios, too, but are treated separately because of their higher relevancy. Finally, complicating factors complete the scenario description by taking into account special circumstances, caused for example by other traffic participants, the road layout or environmental conditions. The overall concept has been presented and examined in an international expert workshop. To handle the issues and open points that have been identified in the workshop, several new scenarios and complicating factors have been added. The basic validity of the concept and the practical realisation in particular is demonstrated by an implementation in form of an automation scenario detection on the traffic dataset 'highD'. The insights gained in the process have been used to further enhance the concept. The final results manifest in a codebook containing the description of all scenarios including their relevant metrics.