Environment Simulation for the Development, Evaluation and Verification of Underlying Algorithms for Automated Driving

Abstract

Algorithms for situation interpretation and prediction in traffic environments are essential in order to face automated driving challenges successfully. At the same time it is hardly possible to guarantee the functionality of such algorithms in every imaginable situation. In order to cope with such problems, the common approach in practice is to apply exhaustive and time consuming evaluation in real traffic situations. In this contribution a driving environment simulation is presented that is suited for the simulation of various highway scenarios and the generation of simulation data comparable to the output of a physical sensor data fusion architecture in real highway situations. The simulation models the dynamics of traffic flow augmented with a data based model to account for perception uncertainties. Finally a sample traffic scenario and the generated simulation output is presented.