Focus area IV: Validation

As the level of automation and connectivity increases, so does the complexity of vehicle validation. A major challenge lies in defining the requirements and the Operational Design Domain (ODD). These form the basis of established safety verification processes for a “pass/fail” assessment of a “System under Test” (SuT). The goal is to demonstrate that the SuT fulfills its intended function within its ODD. However, in reality, specifications are often incomplete. This is because it is impossible to account for all traffic situations in the specification. For this reason, ISO 21448, which addresses the safety of automated vehicles, requires that ongoing risk analysis be conducted even during the operational phase. Uncertainty is thus the central issue in testing. Particularly through the use of AI methods, data- and model-based uncertainties are even accepted.

The goal of the validation focus area is to use the uncertainties arising from perception, environmental prediction, and maneuver planning to improve the ODD and the requirements specification. This lays the foundation for optimizing the validation process through an appropriate specification. The focus is on a scenario-based approach in which a set of relevant traffic scenarios generated from a test model serves as the basis for the validation strategy. Technologically, the testing method is based on model-driven testing. This means that all test case management is carried out via a test model. Although this approach is not yet widely used in the automotive industry, it offers several advantages, including automated test case generation, higher test coverage, reusability, flexibility, and applicability to systems of any complexity.