We present a method for automatically detecting abstract traffic scenarios in live simulations and instantiating them as concrete, testable cases. Scenarios are defined via spatio-temporal and relational constraints between traffic participants, independent of the simulator. Combined with metrics like proximity, criticality, and interactions, the method continuously identifies relevant situations in CARLA simulations. Detected scenarios are stored in the MOSAR database and exported for scenario-based validation, reducing manual effort and improving coverage. This modular, scalable approach enables efficient, realistic testing of autonomous vehicles in complex traffic environments.

• What are abstract scenarios and how their characteristic elements are structured
• How live simulation benefits to scenario databases and system validation
• How to detect concrete scenario occurences in a simulation based from it's abstract description
• How to filter and refine detection results using complementary metric parameters (proximity, criticality, ...)