To safeguard autonomous driving systems, it is necessary to know the operational environment - also referred to as the operational domain in the VVM project - in detail and to be able to analyze it. To this end, two new methods were developed in the project that address the stakeholder concerns with regard to the behavior of an automated vehicle in the operational domain. Two new publications in the project now present these methods.
The first method, the phenomenon-signal model (PSM), is primarily used to formalize the target behavior of an automated vehicle in traffic. This was already presented in the previous newsletter – the new publication (Link below) now presents the PSM in English. The second, directly related method, the Semantic Norm Behavior Analysis, is designed to support the traceable elicitation of behavioral requirements for an automated vehicle in its operational domain. These requirements are formalized by using ontologies as a semi-formal representation of facts and rules that constitute the vehicles target behavior. These facts and rules can be further utilized for a formal behavior specification in the PSM.
Both methods thus serve in particular to analyze the open context, which is one key challenge for automated driving. This associated increased complexity of the environment is taken into account with these methods. Previously existing methods showed the correspondences between the stakeholder concerns and the target behavior of the vehicle rather implicitly - through the two new methods, these can now be more explicitly and traceably argued.
In the further course of the project, these methods will also be used in the final safety argumentation and in the safety assurance of the architectural design and thus have a significant influence on the project results. The two new publications are available under the following links: