
The data availability is a major issue to develop artificial intelligence (AI) tools. On top of that, aside from the availability, the same data entity (i.e. traffic average speed and traffic flow) can be tagged and represented in many different ways. One traffic operator from the city A can capture the average speed and save it in its corresponding data base with the name “averageSpeed” with camel case and in units of “km/h”, while in city B it can be stored in snake case (“average_speed”) and in units of “m/s”, or even using different names (averageTrafficSpeed, traffic_speed_avg) or different languages.
From the traffic management side, the same city can have different traffic operators and several information providers. For a good coordination among all the stakeholders the information must flow smoothly. This requires that the data that are exchanged must have a common structure known by everyone to avoid inconsistencies and errors. On the technological side, data inhomogeneities slows down the development of AI models and the research in all the fields, as extra time is required for both the developer, to understand and adapt the data, and computationally to execute new data transformers.
The adoption of standards is the key point in order to have an efficient data exchange. Standardised data and information allow an efficient exchange at operational level, but it is also required to make any data-based tool more easily reusable in different contexts. This will save time in non-relevant tasks and more efforts can be used in actual research, and also for taking fast management actions from the operator side.

The FRONTIER project has adopted two approaches: on one side the Saref4Auto ontology [1] to identify the entities that coexist in the road mobility sector and the relations among them, and on the other side the Datex-II standard [2] to homogenize those entities and allow an efficient data exchange between the traffic operators to facilitate multistakeholder collaboration. This adoption also allows that tools developed in a certain context can be easily adapted to other different sites. The Figure above represents the workflow used in FRONTIER for data management and exchange. A data provider (i.e. traffic sensor) generates data in its own format. Then a data transformer tool (data harmonisation block) transforms de data according to the standard (Datex-II) and publish them in a context broker to make them accessible to any client that needs access.