Simulation abstrakter Umweltszenarien zur Validierung kamerabasierter Fahrspur- und Elementerkennungsalgorithmen für den Einsatz auf autonomen Modellfahrzeugen


This paper is concerned with improving the testing conditions within the Carolo-Cup project through the addition of a simulator for generating testing images. Within the Carolo project, autonomous model vehicles are developed for participation in the Carolo-Cup at TU Braunschweig. After briefly presenting the Carolo project, parameters such as the competition’s set of rules and the testing conditions at the HAW are presented, including the resulting difficulties for the Carolo project. Subsequently, an overview of different approaches for the testing of embedded systems will be given, including Hardware in the Loop, Software in the Loop and live testing, before presenting current game engines and a robotic simulator which are considered regarding their usefulness for the project. Finally, the Gazebo simulator will be used to show how a simulation environment can be created for the different environmental scenarios included in the Carolo-Cup. The aim of this paper is to advance the development of camera-based recognition algorithms and thus improve testing conditions for autonomous vehicles.