DescriptionTest Case Generation For ADAS Validation Via Cycle GAN Today’s automobile is equipped with a large amount of electronic circuits to achieve intelligent functions, such as collision avoidance, traffic sign detection, etc., for autonomous driving. To meet the safety standard, ensuring extremely small failure probability over all possible operation conditions is one of the critical tasks for an autonomous driving system. However, physically observing all these corner cases over a long time is almost impossible in practice. In this project, we use machine learning algorithms to efficiently generate corner cases that are not easy to observe.
OrganizationDuke Unversity
DepartmentElectrical and Computer Engineering
Sponsor Campus GridOSG Connect
Principal Investigator
Xin Li
Field Of ScienceEngineering