Francesco Marchioni, Alberto Fontebasso and team

The SMART project provides innovative traffic management software and supports smart urban mobility. The system combines the immediacy and precision of Galileo positioning data with the power and predictive capabilities of artificial intelligence. The SMART platform can be deployed on public road networks for immediate and autonomous monitoring in order to promote traffic reduction and road safety. Through a dedicated recurrent neural network model and a deep reinforcement learning approach, the platform is capable of detecting and predicting traffic patterns from positioning data shared by smartphones and vehicles. It then optimises traffic signals autonomously and reduces traffic and pollution by interfacing with the local traffic management centre.
Moreover, a database will be produced from the shared positioning data to obtain insightful information on spatial and temporal evolutions in traffic density, speed, and transit times. This information will be beneficial in urban planning, policy making, and resource allocation.