The field of artificial intelligence (AI) is constantly evolving, and autonomous vehicles are a striking example of this. For these vehicles to be able to navigate roads safely, it is crucial that they can learn to identify and react to the different situations they may encounter. This requires a significant amount of data collected from real-world situations or simulations.

To accelerate the development of autonomous vehicles and improve their safety, IMREDD has launched the BonsAI project. This is a testbed, i.e. a reduced-scale road environment.

The originality of this project lies in the fact that the test environment reproduces a real location: the Balcons d’Azur roundabout in Mandelieu-La Napoule. This roundabout has been equipped with sensors to collect real and dynamic data on road traffic. The data collected is sent to the testbed and used for
to test and develop autonomous driving technologies, evaluate the performance of navigation and communication systems, and improve the safety of autonomous vehicles.

Scale-model vehicles, equipped with the same technology as autonomous shuttles, are operating on a scale model of the roundabout at the IMREDD Technology Platform and responding to traffic conditions detected by sensors installed at the roundabout in Naples.

The BonsAI project also aims to contribute to the development of a new machine learning approach called «federated AI». Unlike
The traditional approach, which involves collecting individual data in a centralised database, federated AI is based on distributed learning, where data remains on local devices and data models are exchanged rather than individual data. Thus, federated AI allows for faster and more efficient learning while preserving the privacy of individual data.