Problem statement

Map and navigation services represent an indispensable foundation for many innovative applications in the vehicle. In addition to static road information such as traffic signs and traffic lights, dynamic events e.g. traffic jams and construction sites are assigned to the map data. Numerous background data can be processed by the vehicle and therefore support the optimization of the energy consumption or improve the active safety. The feeding of current data on the quality of the road surface can further improve road safety and driving comfort. At the moment, the quality of the road surface of motorways or federal roads is being recorded alternately at intervals of 4 years with special measuring vehicles. However, with the current system, the acquisition of the corresponding data is very complex because of the large road network. In addition, the information is only available on a delayed basis.


The aim of the project "QoStreet" is to develop and validate a method for the classification of the road surface based on smartphone sensory data. The surface type should be determined as well as the quality of the road. This is to be done with the help of smartphone sensor data, which were recorded in the course of numerous fleet tests at the Institute of Automotive Engineering. In particular, weather data in the context of the BMVI provide additional information, which are necessary in the classification process.

The results can be used to increase the driving safety and the driving comfort. In addition, the maintenance management of the road construction department can be specifically supported and made more comprehensive. The effort and costs for determining the surface quality of roads can also be potentially reduced compared to the use of conventional measuring vehicles.


Initially, anonymously collected sensory data are assigned to the travelled road sections and calibrated. The data from existing fleet tests are merged with weather data from the German Weather Service to take account of changing environmental conditions and to make the data quality plausible. The large amount of sensor data is then used to classify the road quality in the area around Munich. The algorithm designed to calculate the road surface is checked by means of conventional comparative measurements. After project completion, the generated metadata are to be provided via a data portal.