Testing is an integral part of ensuring safety for automated driving systems. Moving away from distance-based approaches, scenario-based development has become the industry's state of the art. Therein, concrete test cases are generated from scenario classes that should cover the operational design domain. The number of scenario classes is finite, but each scenario class describes an infinite number of concrete test cases. For a given scenario class, test suite quality depends on the number of test cases and the value of these test cases. In this talk, we explore how test suite quality can be assessed appropriately.

• What are the problems in testing automated driving systems?
• How does scenario-based testing work?
• What constitutes a good test suite?
• How can test suite quality be assessed quantitatively?
• What are the preliminary experimental results for quality functions?