Dr. Radusch, you recently gave a presentation at the industry meeting of the “Verband Deutscher Auskunfts- und Verzeichnismedien” in Berlin. What do directories like Das Telefonbuch have to do with mobility?

Directory data is interesting when it comes to so-called “points of interest,” i.e., locations relevant to mobility services. Together with pdm solutions, the company behind Das Telefonbuch, we evaluated how such data can be used. Specifically, we integrated curated location data for kindergartens, playgrounds, and schools into a driver assistance system that warns drivers about particularly vulnerable road users. The requirements for data quality are especially high in this area. We examined how such information can be combined with other mobility data and integrated into the Digital Twin. In the future, this will enable not only safety applications but also classic telematics services, such as barrier-free routes to the nearest pharmacy or themed city tours.

What other data sources feed into the Digital Twin?

To create a realistic Digital Twin, we bring together a wide variety of data sources: information from traffic management centers, stationary sensors, such as those in traffic light systems, as well as sensor data from vehicle fleets, including highly automated vehicles. Instead of relying exclusively on specialized measurement vehicles, we also use crowdsourcing. With our AI-based app Eidos Road Glancr, a standard smartphone mounted on the windshield can collect information about the road environment in compliance with data protection regulations. We have already successfully tested the app in field trials with buses and taxis. For rare or safety-critical events that rarely occur in real-world data, we supplement this data with simulations in our open-source environment Eclipse MOSAIC.

How do you manage to integrate such diverse data?

A key goal is to break down existing data silos. This also supports digitally sovereign systems, as not every data provider has to meet all quality requirements of all, often still unknown, data users from the outset. The requirements vary greatly: While correct addresses are usually sufficient for directory data such as Das Telefonbuch, vehicles and driver-assistance systems require highly precise geocoordinates. One example is schools or playgrounds, which often span entire city blocks and border on multiple streets. If an assistance system were to issue warnings across all adjacent streets, this could lead to a habituation effect, with negative consequences for traffic safety. That is why such information must be processed in a targeted and precise manner so that warnings are issued only where there is actually an increased risk. Here, our expertise in the automotive sector and our AI systems help us to efficiently integrate and validate this within the Digital Twin.

What are the biggest challenges right now?

As just described, clearly data quality: data must be up-to-date, precise, and reliable. For example, an announced construction site does not necessarily mean that it will actually affect traffic or that it will be implemented exactly at the planned location. This is where our previously mentioned Road Glancr app comes in: It automatically detects changes in the road environment and evaluates them, such as the extent of construction sites, new traffic signs, road damage, or available parking spaces.

What role does artificial intelligence play in this?

Artificial intelligence supports us in several areas: in the automated detection of relevant objects in the road environment, in reconciling conflicting data sources, in identifying outliers, and in making predictions. This allows a digital twin to not only represent the current state but also predict developments, such as traffic or parking conditions. Anyone involved in developing digital road maps knows how unique different locations and conditions are. Only with the help of AI can we efficiently analyze and correctly classify all these different situations.

How can the Digital City Twin be utilized by cities and the mobility industry?

For cities, a Digital Twin lays the foundation for data-driven decisions toward sustainable and user-centered mobility that takes all road users into account. Through open, standards-based interfaces, we offer mobility service providers the opportunity to develop and deliver innovative services for different target groups.