How can artificial intelligence replace vehicle counting sensors on the carriageway in order to help road operators manage traffic more effectively and save money?
The Île-de-France Highways Department (DiRIF) is the public body responsible for motorways and trunk roads in the Île-de-France region. In order to provide real-time user information and to carry out technical research to improve traffic flows, it draws on vehicle counting data captured by sensors installed under the carriageway. These sensors are now ageing and will need to be replaced. The total investment required would be in excess of tens of millions of euros and the work is unlikely to be carried out in its entirety.
maintenance bill for DiRIF sensors
in the coming years
The DiRIF has access to large amounts of historic counting data captured by its sensors. It also has GPS speed data, which was purchased at a significantly lower price. With the emergence of Big Data technologies and data science, this data offers a fresh opportunity for the DiRIF to take a more innovative and effective approach to its operations. The challenge is to study the possibility of exploiting the hundreds of millions of items of historic data with the aim of developing an artificial intelligence algorithm to predict counting data and thus avoid the necessity of replacing physical sensors.
Wintics develops AI solutions for Smart Cites. These solutions focus on three themes: mobility, urban infrastructure and property management. Wintics has demonstrated the effectiveness of its algorithms on a number of occasions by winning international competitions relating to topics such as predictive maintenance of transport infrastructures and identifying excess energy consumption in buildings.
The algorithm developed by Wintics during the challenge can model vehicle numbers at 6-minute intervals over the entire experimentation zone (250km of expressways) with 95% accuracy, using just 20% of the existing sensors. If this solution was scaled up across the entire Île-de-France network, it would improve data availability and current service levels, while also saving tens of millions of euros in sensor replacement costs. DiRIF is already working with Wintics to integrate the algorithm into its information systems so that the reconstructed vehicle counting data can be made available in real-time to operators and users.