Artificial Intelligence to improve reliability of autonomous mobility

The technology center Ceit participates in the project AutoTrust, an initiative focused on the development of reliable functionalities based on Artificial Intelligence (AI) for the field of cooperative, connected and automated mobility .

Ceit is working on the detection of interferences in road infrastructures caused by vegetation growth trying to explain the AI results through Explainable Artificial Intelligence techniques.

In the complex field of cooperative, connected and automated mobility (CCAM), there is still a large reliability gap, one of the pillars of autonomous driving. The technology center Ceit, is one of eight partners core topic that is participating in the project AutoTrust, whose goal is to propose practical Artificial Intelligence (AI) methodologies that provide reliability, transparency and safety to AI systems.

A danger that often arises for both human drivers and in-vehicle artificial intelligence systems is the blockage of traffic signs or other essential elements of road infrastructure caused by the growth of vegetation near roads.

Interference identification 

In this sense, Ceit is working on this issue by applying Explainable Artificial Intelligence (XAI) techniques, which allow not only to identify these interferences accurately, but also to understand and explain how AI systems detect and respond to these situations.

XAI techniques provide information that allows understanding and explaining the results provided by intelligent driving algorithms when faced with complex situations such as occlusions and in scenarios where vegetation can compromise road safety. For developers, this means better understanding how the algorithms work by providing detailed information. For end users, such as those in charge of maintaining road infrastructure, it provides understandable information that facilitates decision making.

This ability to interpret and explain strengthens the confidence of users and regulators in emerging technologies, thereby improving road safety for the benefit of drivers and autonomous vehicles.

AutoTrust is supported by the Basque Government through the ELKARTEK call for 2023. At partnership with Ceit, and under the coordination of Vicomtech, the consortium is formed by University of the Basque Country (UPV), Automotive Intelligence Center (AIC), Basque Center for Applied Mathematics (BCAM), Tecnalia, Mondragon Unibertsitatea and Ikerlan.