The Agency for Roads and Traffic (AWV) is responsible for fighting smoothness on Flemish roads under its own management. To this end, AWV's winter service is supported by various data sources and technical tools.

AWV has been investing for some time in its own smoothness measurement system (GMS) that combines real-time results from measuring points and measuring stations with weather forecasts from the KMI.

Using this info, district employees independently take decisions whether or not to grit in their district. In each gritting decision, they weigh up road safety and environmental impact.

The knowledge and skills of the employees are very different and translate into inconsistent gritting policies across districts. However, roads do not stop at the boundary line between two districts.

AWV therefore wants to develop a system that can support district employees in their gritting choices and lead to more uniform gritting decisions across districts.

By combining physical rules and machine learning techniques, the system can help weigh up and indicate which type of gritting agent in what quantity and at what time is most appropriate to make and keep roads in the district free of ice.

At the same time, it allows an automated system to track and learn from the scattering decisions. This could be useful for knowledge transfer between staff members, and other training purposes.

The project was aptly named 'Elsa' after the main character of the popular animated film 'Frozen'.