Classification of Driving Maneuvers in Urban Traffic for Parametrization of Test Scenarios

This contribution proposes a semantic description of vehicle behavior in urban environments in the form of maneuvers. By detecting these maneuvers in recorded measurement data of test drives, specific scenarios can be identified and evaluated in regards to vehicle behavior on public roads. Characteristicsofextractedmaneuverscantheninturnbeused to describe and differentiate the behavior of traffic participants in test scenarios for automated driving systems. We extend an earlier concept to be applicable for urban vehicular traffic. Furthermore we analyze real-world measurement data of test drives by identifying the maneuvers which are defined in this contribution. We then show exemplary how classified vehicle behaviorcanbeusedtoformulateaLogicalScenariocontaining parameter distributions. These distributions are then sampled to obtain multiple Concrete Scenarios for use in simulation. Finally, we discuss our results and possible future work.      «

This contribution proposes a semantic description of vehicle behavior in urban environments in the form of maneuvers. By detecting these maneuvers in recorded measurement data of test drives, specific scenarios can be identified and evaluated in regards to vehicle behavior on public roads. Characteristicsofextractedmaneuverscantheninturnbeused to describe and differentiate the behavior of traffic participants in test scenarios for automated driving systems. We extend an earlier concept to be applic...     »