Untersuchungen zur Anwendung von computergenerierten Kamerabildern für die Entwicklung und den Test von Fahrerassistenzsystemen (Volume 88) (English shop)

Preview

Extract, PDF (770 KB)
Table of Contents, PDF (120 KB)

This dissertation develops methods for testing camera-based driver assistance and active safety systems in simulation. The focus of the application of these methods is the hardware-in-the-loop testing of networked driver assistance systems. This use case provides for an end-to-end test of the functions, i.e. the sensor data should be fed in as close to the beginning of the signal processing chain as possible. In the case where a camera system is to be stimulated, the generation of a synthetic camera image by means of computer graphics presents itself as a suitable approach. Given the constraint that a model is always a simplification of reality, it is important to investigate what degree of realism the simulation must exhibit in order to plausibly stimulate the image processing. Within the scope of this work, a comprehensive validation of computer-generated camera images is carried out. On the basis of these findings, possible fields of application for simulation in testing and development are defined. To strengthen the methodological work in computer graphics, a hardware-in-the-loop test bench for testing driver assistance and active safety systems was developed at AUDI AG. These systems have been used to accompany development work since 2010. The experience gained from their application made it possible to strengthen the methodological development.