The role of AI is increasing in every sector. One of them is road maintenance. At HTWK Leipzig and FTZ Leipzig, geotechnical specialists and electrical engineers are developing an AI-driven sensor system designed to monitor real traffic loads and support the prediction of road renovation needs.
Increasing need for accurate traffic load data
According to a study by the German Institute for Urban Studies commissioned by the German Construction Industry Association, roads in many federal states and municipalities are in need of repair. More than €370 billion in investment will be needed by 2030 to maintain and expand the rail network, roads and paths in Germany.
New approach to monitor road infrastructure
The project RoadIT1.0, launched in 2023 and funded by the Federal Ministry of Transport until the end of 2025, aims to create a cost-efficient and highly robust measurement system for widespread use across municipal and regional roads. The concept relies on compact sensors embedded directly into the road structure. These sensors capture a wide range of physical data, while AI algorithms interpret the information to determine the actual loads exerted by passing vehicles.
‘’The sensors provide us with data on the number and weight of vehicles travelling on a road. The AI uses this information to calculate the actual traffic load. This has been rising steadily for years, and both the number and size of vehicles are reaching levels for which most roads were not originally designed,’’ says Ralf Thiele, professor of geotechnical engineering and head of the research project.
Functional testing began in spring 2025 on a company-owned road in Frohburg near Leipzig, where several sensor prototypes were evaluated over a six-month period.
Data quality and AI-based interpretation
The test installation was successful: the system started delivering continuous measurements.
The next step involved installation on a public road near Leipzig, in Oelzschau, to verify how the technology performs under uncontrolled, everyday traffic. There are currently three measurement lines, each with five sensors. Each sensor generates thousands of data points per second, offering detailed insights into the forces exerted by passing vehicles.
By combining multiple measurement principles, the system identifies axle loads, axle spacing, vehicle speed, and vehicle class with high accuracy. Weight estimations achieve a relative precision of less than one tonne, and axle spacing can be determined to within approximately ten centimetres.
The volume and complexity of the data require advanced processing. Alongside analytical methods, the project relies heavily on machine learning to detect patterns that are difficult to interpret with conventional techniques. This approach increases robustness and enables the system to adapt to different road structures.
The data is displayed in real time on a web dashboard on the screen. The geotechnical engineers receive support from other project partners: Infratest Digital Solutions transfers the data from the sensors to the dashboard, and N4 Leipzig takes care of visualisation and real-time display.
Towards predictive and sustainable road maintenance
The project's goal is to provide municipalities with an affordable, scalable monitoring solution. With continuous data on actual traffic loads, maintenance teams can transition from reactive repairs to predictive planning. This change could improve safety, extend service life and reduce long-term infrastructure costs.
/White%20Versions/dehub_white.png?width=115&name=dehub_white.png)
/White%20Versions/stadt_leipzig_white.png?width=130&name=stadt_leipzig_white.png)
/White%20Versions/sachsen_signet_white.png?width=90&name=sachsen_signet_white.png)