i2CAT will be at SCEWC showing a traffic risks and flows detection system based on AI and big data and an advanced cybersecurity solution for autonomous vehicles.
Artificial intelligence to detect traffic risks
Within an initiative of the Centre of Innovation for Data tech and Artificial Intelligence (CIDAI), the i2CAT Foundation has led the development of a system to detect risks and traffic flows on urban and interurban roads based on Artificial Intelligence and big data. The research is a collaboration between Barcelona Supercomputing Center, the Centre de Telecomunicacions i Tecnologies de la Informació (CTTI), Eurecat and Microsoft. It aims to proactively prevent traffic accidents by identifying risk situations and analysing traffic flows to detect anomalous behavior and reduce traffic jams.
Within the project, i2CAT has developed a video processing system based on deep learning algorithms and computer vision that identifies all the objects on the road (including vehicles and pedestrians) and extracts information related to their position, trajectory and speed. To identify risk situations, i2CAT has designed a semantic corpus including specific road information (lanes, traffic directions, traffic lights, etc.) and a series of regulations that describe different risk situations according to the type of road.
The system has been tested in two segments of the C31 road and in an urban intersection in Barcelona. On interurban roads, it identifies risks such as unauthorized lane changes or driving in the opposite direction. On urban roads, it identifies threats such as, for example, pedestrians crossing at a red light or vehicles that get too close to pedestrians.
The system has demonstrated a success rate of over 83% in different situations and environments.
“The traffic risks and flows solution developed by i2CAT is an opportunity for the administration to act preventively and to reduce traffic accidents and deaths on the road. Additionally, it could help traffic flows become more efficient and, thus, improve air quality”, explains Àngel Martín, AI Innovation Manager of Distributed Artificial Intelligence at i2CAT.
Security and privacy in a connected environment
Within the European project CARAMEL (Artificial Intelligence-based cybersecurity for connected and automated vehicles), the i2CAT Foundation has led the implementation of a system of vehicle-to-vehicle and infrastructure communication protocols (V2X) that targets the detection and reduction of attacks that an autonomous vehicle could receive and that could compromise its security.
The system, developed from open-source software, runs in On-board Units (OBUs) and communication units located on the side of the road, known as Roadside Units (RSUs).
Its purpose is to authenticate the origin of the communication packets transmitted between the different vehicle-infrastructure units (OBUs and RSUs), and detect and stop possible V2X message transmission attacks. An additional layer of privacy has also been incorporated, consisting of an artificial intelligence algorithm that decides to change the signature every time, thus avoiding the malicious tracking of vehicles that could compromise users’ privacy.
“Vehicles are increasingly resembling data centers on wheels and, therefore, are increasingly vulnerable to attack. CARAMEL’s results provide a very significant advance in terms of cybersecurity for the next generation of autonomous and connected vehicles” explains Jordi Guijarro, project coordinator and director of Innovation in Cybersecurity at i2CAT.
This software will allow companies in the IT sector to develop different applications. For instance, the location of the vehicle by GPS will offer an alternative to toll control systems by precisely calculating the vehicle’s trajectory and, therefore, requesting automatic payment on specific highways. i2CAT has already tested this solution in collaboration with the Catalan company Tecsidel.
These and other projects and initiatives that the i2CAT Foundation is working will be on display during the SCEWC at the Government of Catalonia Pavilion (Hall 2, E61, Booth 16).