Passive testing of highly automated driving systemsŠelih, Luka (Avtor) Rodič, Miran (Mentor) Župerl, Uroš (Mentor) Dodig, Dino (Komentor) autonomous drivingtestingsafetyCARLAAEBDriver assistance systems and automated driving have become one of the most important components of new vehicles. Enormous progress has been made with developing algorithms for autonomous control and perception, but the problems of safety validation remain. In this master thesis, we first briefly summarise safety and testing approaches in autonomous driving. Then we propose our own approach for testing automated/autonomous driving, which we call "Passive parallel Testing". The approach focuses on detecting discrepancies between the human driver and the control of the automated driving function, thus identifying potentially critical situations. The test method is demonstrated at the end in a simulation with the CARLA simulator.[L. Šelih]20212021-09-08 14:31:57Magistrsko delo/naloga80350sl