Multi-scale modelling for smarter, greener traffic management
An EU-funded project is laying the foundations for next-generation urban traffic management systems that are capable of intelligently anticipating congestion, rather than simply reacting to it. Using innovative modelling techniques and smart city technologies, the research promises to make trips faster while reducing urban transports environmental impact.
© LICIT (IFSTTAR/ENTPE)
The twin issues of congestion and climate change mean it is essential to get people out of their cars and on to public transport. In theory, encouraging more people to use buses rather than their own cars to commute to work should ease congestion by reducing the number of vehicles on the road. But as roads become less clogged, the incentive to switch diminishes, so more people use their cars and traffic congestion continues.
What if the trips that contribute most to congestion could be avoided dynamically at specific times and on certain routes, and the most efficient mode of transport selected by individuals across the entire city?
The team leading the EU-funded MAGNUM project is adopting a fundamentally new strategy to address the challenges involved in the day-to-day management of urban mobility. Instead of looking at the flow of traffic in bulk along specific routes or between city districts, the researchers are focusing on individual trips to develop dynamic, predictive and accurate multi-scale traffic models for a whole city.
The ultimate goal is to rethink the daily management of mobility in smart cities, says MAGNUM coordinator Ludovic Leclercq of IFSTTAR, the French Institute of Science and Technology for Transport, Planning and Networks. Advances in technologies provide new options for more efficient transportation systems, but a key characteristic of dynamic transportation networks is the classical antagonism between the user and the system: the integration of best choices for users does not necessarily lead to optimal conditions for the system.
More realistic models
Focusing on the bigger picture while keeping track of individual behaviour poses notable challenges. These include the need to ensure consistency in traffic patterns at local and global scales, and coping with the heterogeneity of individual trip choices caused by different departure times, travel speeds and route decisions.
To address such issues, the MAGNUM researchers are developing a modelling framework for predicting traffic that takes into account more realistic behavioural models.
They are also investigating new traffic data-analysis methods and have realised a major breakthrough at large urban scales by defining the concept of 3D congestion maps, which provide a unique picture of the daily pulse of traffic flows and congestion.
The rapid deployment of new products and services, such as autonomous vehicles and smart ride-sharing applications, are set to have a profound impact on how individuals move around cities.
The MAGNUM models are being developed with those trends in mind, with a view to helping individuals, companies and cities anticipate the potential drawbacks of novel transport methods while harnessing their benefits for the efficiency and sustainability of urban transport networks.
A first application was to design a completely new travel-time estimation method that works in real-time based on matching current traffic conditions with the closest representative congestion maps.
Support for urban planners
The research has also been supported by the development of a unique massive multi-player simulation game platform that enables the researchers to study in a controlled environment how users react to traffic information and decide their routes and transportation options. The current test case is the French metropolitan area of Lyon, which supports more than 1 million journeys per day. The game enables researchers to create refined behavioural and traffic models by monitoring player responses and interactions.
Traffic congestion has negative social and economic impacts. We aim to make the transportation system more efficient but also more sustainable as environmental effects will be taken into account when optimising the models, Leclercq says. The results are impressive, especially in light of the relative simplicity of the method.
The researchers plan to make their models available under an open-source licence at the end of the project, allowing companies, urban planners and city transport officials to develop improved traffic management systems, policies and regulations.