The purpose of this dissertation is the development of a system for autonomous driving testing in a simulation environment. We selected the software tool SCANeR DT, which enables the generation of different scenarios and critical situations and also includes the implementation of sensors, which are an essential part of autonomous driving. We described these sensors in detail and also compared them to the simulated ones inside the simulation environment. Additionally we developed customized modules, which are necessary for the autonomous vehicle to function and also two algorithms for the testing image processing of images from the camera. We tested our autonomous vehicle on a predefined route which consisted of multiple critical situations. We compared the results from our autonomous vehicle with results from human drivers, which drove on the same route. The results show us that our autonomous vehicle got better scores, which means that it performed better in critical situation than human drivers.
|
