Sinan Hasirlioglu,
"A Novel Method for Simulation-based Testing and Validation of Automotive Surround Sensors under Adverse Weather Conditions"
, 2-2020
Original Titel:
A Novel Method for Simulation-based Testing and Validation of Automotive Surround Sensors under Adverse Weather Conditions
Sprache des Titels:
Englisch
Original Kurzfassung:
Equipping vehicles with driving automation systems aims to increase the driving comfort and eliminate the human error that is responsible for the majority of accidents. Automation systems typically use surround sensors to monitor the driving environment. Recently, accidents with automated vehicles have shown that errors in sensor data measurement and interpretation can lead to fatal injuries. It is thus necessary to test the reliability of environmental perception systems before their market introduction. In practice, real-world testing is expensive, time-consuming, and not reproducible. Furthermore, the increasing system complexity generally leads to an increased test effort and, as a consequence, longer development time. To counteract this, test facilities and simulation will play a central role in the future process of development, testing, and validation of driving automation systems. Adverse weather conditions such as rain and fog usually occur randomly, however, they degrade a sensors performance due to absorption and scattering processes. Therefore, environmental conditions need to be considered to ensure safe automated driving. This thesis presents a general approach for simulating the effects of adverse weather on raw sensor data in real and virtual environments. The focus is on the influence of rain and fog on data from camera, lidar, and radar sensors. Further, a novel test method is proposed that uses both the real and virtual simulation approach to minimize the effort required to safeguard environmental perception systems under adverse weather conditions. (...)