On the Way to the Real Data Highway
Today, if you want to know what the structural condition of a road is, you have to take some cores from the road pavement, analyse the material and deduce the overall condition. “Road closures are often necessary for this, and the procedure also weakens the road structure“, explains Ivan Isailović, Group Manager for Road Construction International at the STRABAG Competence Centre TPA in Cologne. He is involved in a project in which the TPA is researching, along with the BASt and the Data Science Subdivision of STRABAG Innovation & Digitalisation (SID), how to directly obtain the pavement structural condition based on sensor measurements.
In the “Relaxed Hybrid“ research project, the TPA has installed 40 sensors in a hybrid road pavement. This construction method corresponds to a concrete pavement and asphalt layers built on top of it. It’s “relaxed“ because the concrete layer is first broken up into slab sections to remove so-called constraint stresses.
The parameters measured by the sensors in the project, such as acceleration and temperature, should help to determine how the whole system behaves, including the relaxed concrete layer. So far the design of the hybrid construction has been based mainly on assumptions – experts strongly suspect that the thickness of the hybrid road pavement is currently over-dimensioned, which generates unnecessary costs. The measured values of the installed sensors have shown that the “relaxed hybrid“ exhibits significantly higher bearing capacity than originally thought.
Material parameters are calculated automatically from the sensor data using machine learning. The potential of such an application goes far beyond planning the right thickness of the hybrid construction: Monitoring the condition of the material and the loads would also indicate damage in the road before it becomes visible as a pothole.

3.4 million axle passes – the experimental vehicle on the test track of the Federal Highway Research Institute (BASt) simulates million-fold passes of 10-tonners over the asphalt. And everything is well documented by sensors from STRABAG in the pavement.
Huge amounts of data
The 40 sensors have generated huge amounts of data since 2019. “Processing this efficiently was difficult for us at first“, says Isailović. Sensors also generate errors and processing them requires special know-how. The TPA found this know-how within the company Clarify Data – the start-up took over the data analysis, accompanied by the experts of the SID Data Science Subdivision, in the Vienna team of Gerhard Höfinger, Function Lead Data and Decision Analytics. “Clarify Data is a very competent partner, so our role was not to control their results, but to understand them in detail so that we can reliably apply this solution to other projects“, explains the data specialist.
Pioneering work has been done on both sides of this project. The experts arrived at comparable results to conventional methods when evaluating the sensor data. Recommendations for the sensors to be used in the future and their placement can now be derived from the experience gained. The next project is already in the planning stage: it will now go on the real road. “Especially for our public–private partnership (PPP) projects, sensor data would be a very exciting proposition. We commit ourselves to maintain the road for 30 years in most cases. We could derive measures from the sensor data early enough and avoid major damage“, says Isailović. He is already thinking ahead. “Perhaps where today a roadside emergency telephone is installed, there will soon be a physical data hub that can be used to communicate with the outside world and directly with the road.“
Gerhard Höfinger concludes,
“We can only implement projects within the framework of the digital strategy together with the operational units. It was very interesting to work with a start-up, to see what tools and software they use.“ Höfinger also sees the project as confirmation that STRABAG is on the right track regarding data science. He adds that the cooperation with TPA will be further expanded: “We are now looking for suitable databases for bitumen investigations.“