Newsletter No. 3

Mid-term Event: Questions and Answers from the Event

Advances in the use of the phenomenon-signal model in VVM

With the help of the Phenomenon Signal Model (PSM), information flows of traffic scenarios are being investigated in the project. After discussing the basic questions of the model last year, the VVM consortium can now approach possible applications and evaluations. The next steps will be computer experiments with the corresponding software.

The PSM considers the flow of information in given scenarios. It examines which information is sent between agents and which information the ego vehicle must receive and decode in order to move safely through the scenario. In other words, what facts - i.e., circumstances in the scenario - must the ego vehicle recognize? In the context of PSM, Semantic Norm Behavior Analysis is a structured approach to identify such relevant facts based on diverse sources. Such sources - sources of norm behavior - can be laws, accident reports or expert knowledge. In the project, it has now been possible to adapt the method in the last few weeks so that it can also include further elements from VVM such as ontologies and criticality phenomena.

However, the facts are only one important component. The other is the specifications of what the ego vehicle should do with these facts, i.e., how it should behave. This question is also addressed in the Semantic Norm Behavior Analysis. As a result, rules are established that are formalized in the manner of "IF facts THEN action".

These results of the semantic norm behavior analysis - i.e. the set of facts and related rules for a scenario - can now be formalized and executed on the computer. First computer experiments already show various possibilities to evaluate the result of these experiments - a graph which shows the development of facts in a scenario - according to different aspects. The next steps in the project will be dedicated to this evaluation. In this way, it can be investigated which perception is absolutely necessary in order to reliably recognize a situation, which is, for example, a basic prerequisite for braking.

In order to be able to use the PSM formalism, the travel paths of the agents and the ego in a scenario are divided into small pieces - so-called zones. The effect of the choice of such zones as well as their criteria is also an important question that will be investigated.

Therefore, the “PSM analysis” involves both the semantic norm behavior analysis and the formalism with the graph. The entirety of this method is now at a working level in the project in order to close the link to hazard and risk at the behavioral level, but also close the workflow towards capability-based architecture on a solid basis.