Automated scenario generation and iterative regression testing method for autonomous driving systems

Abstract

In this paper, we propose a new automatic test and evaluation method that can be used for autonomous driving systems in virtual simulation test. Aiming at the problems that the existing automated test method cannot guarantee coverage and validity of the test scenarios, and evaluation lacks the consideration of the complete development life cycle of the systems, we firstly introduce a new scenarios generation method that can generate scenarios with higher overall effectivity on the premise of ensuring coverage. Then, an integrated test platform is constructed, in which the components can be automatically invoked by test scripts. Finally, the automated iterative regression testing is used to ensure the effectiveness of the whole development and test process. The application and verification of a level three (L3) Traffic Jam Pilot (TJP) system is carried out on the above-stated method. The result shows that the proposed method can effectively detect system defects and significantly improve test efficiency.