Background: The Operational Design Domain (ODD) of an automated driving function defines on which roads and under which environmental conditions the function is safe to operate. It plays an important role in definition, safety analysis and validation of automated driving. In many cases, users want to determine metrics about ODDs, or about ODDs in combination with other work products, like collections of validation scenarios. Such metrics could answer questions such as what percentage of the road network of a given region is inside the ODD. While language formats to specify ODDs have emerged over the last few years, a solid methodology on how to calculate different sorts of metrics is still ONThe roadmap for the future. Methods: This contribution suggests metrics for ODDs that are mathematically built upon a notion of ontologies, and ODDs as multi-dimensional cross-products of sets, using standard arithmetic and set operations. To illustrate the idea, a couple of possible metrics for ODDs are derived as examples and discussed in the light of some real-world use cases. Results: To illustrate the application of a ODD metric, we apply an analysis of a sample trip and calculate the theoretical availability of variants of an automated driving system with different ODDs. Conclusions: The metrics presented and the shown sample application present an important next step in discussions around ODDs of Automated Driving Systems. They make it possible to not only consider an ODD specification as a reference for a single system, but allow comparing systems with different ODDs, judging the maturity of a system with a certain ODD, or provide indicators how usable a system is within a real-word application.
Keywords: Automated Driving; Operational Design Domain; Scenario-based testing.
Copyright: © 2023 Kaiser B et al.