A system for data-based decisions on cost-effective road maintenance, DatEnKoSt for short.


The basis of the system is low-cost cell phone data, which is used to train an artificial intelligence.


In the future, the system will make it possible to provide an area-wide representation of the condition of the traffic routes and a forecast as well as measures for cost-effective maintenance.

Your contact

Portrait Mitarbeiter Sebastian Kreibich

Sebastian Kreibich

eMail: sebastian.kreibich@aws-institut.de
Phone: +49 681 96777 747

In a nutshell


The research project “Data-based decisions for cost-effective road maintenance”, or DatEnKoSt for short, is developing a low-cost and reliable recording of the condition of traffic routes. Based on this, artificial intelligence will enable a forecast and simple derivation of optimal measures. In this way, unnecessary follow-up costs are avoided. Municipal maintenance of traffic routes thus becomes more efficient and sustainable.


Modern smartphones have numerous sensors that record a wide range of data streams when driving on roads, such as acceleration and braking behavior or image data. This data, which can be captured inexpensively, is recorded on defined roads in the research project and then correlated with the results from measurement vehicles. Procedures already commonly used by engineering companies will be automated with the help of AI. In this way, an artificial intelligence will be trained to provide condition mapping of road conditions in the future using the low-cost cell phone data. In addition, the system is provided with further information about the previous course, other influences such as weather and structural conditions. This data is used to create a forecast that allows the system to simulate the effects of changes to the aforementioned influences. This enables the trained artificial intelligence to derive measures for long-term cost reduction.

initial situation

Prozesskette des Projekts DatEnKoSt

The maintenance of traffic routes is a noticeable challenge. Due to a lack of financial resources at the state and local levels, maintenance can neither be planned in a timely manner nor with foresight. Engineering offices have to generate costly data and image material in order to derive maintenance measures on this basis.

Ultimately, this leads to stress and danger situations, accident costs and delivery delays. What is needed, therefore, are particularly favorable methods that record and evaluate the condition of the infrastructure and derive optimal measures from it.


Condition detection

Measures based on


The August-Wilhelm Scheer Institute elaborates 2 crucial components:

Sensor technology. The transmitted data, such as images and sensors, are analyzed and classified by us. In this way, an artificial intelligence is trained to provide a status evaluation of the current traffic routes from simple cell phone data.


Prediction. Supervised learning is used to develop functions and algorithms with training data where the output is already known. In this way, the system is trained to enable future predictions.

Press releases



Logo Bundesministerium für Verkehr und digitale Infrastruktur

The DatEnKoSt project is funded by the German Federal Ministry of Transport and Digital Infrastructure.

The DatEnKoSt project is funded by the mFund research initiative.