AI for improving reliability of autonomous mobility

The Ceit Technology Centre is taking part in the AutoTrust project – an initiative that focuses on the development of AI-based reliable functions for use in the field of cooperative, connected and automated mobility.

Ceit is working on the detection of interference in road infrastructures caused by the growth in vegetation, in an attempt to explain AI results via explainable artificial intelligence (XAI) techniques.

A reliability gap still exists within the complex field of cooperative, connected and automated mobility (CCAM), this being one of the pillars of autonomous driving. The Ceit Technology Centre is one of the eight key partners that are taking part in the AutoTrust project, the aim of which is to put forward practical AI methodologies with a view to providing AI systems with reliability, transparency and safety.

One danger that often arises both for human drivers and AI systems in vehicles is when traffic signs or other essential elements on roads become blocked as a result of the growth in close proximity.

Taking this into account, Ceit is working on the problem by applying explainable artificial (XAI) techniques to enable not only such interference to be accurately identified, but also how AI systems detect and respond to these situations to be understood and explained.

XAI techniques provide information that enables the results offered by intelligent driving algorithms to be understood and explained when complex situations such as occlusions are being faced and in scenarios where vegetation may compromise road safety. For those developing it, this means being able to understand better how algorithms work by providing detailed information. For its part, it offers end users such as those in charge of road infrastructure maintenance with comprehensible information to help them with decision-making.

This interpretation and explanation capacity reinforces trust on the part of users and regulators in emerging technologies, thus improving road safety for the benefit of both drivers and autonomous vehicles.

AutoTrust is backed by the Basque Government via the 2023 ELKARTEK Programme. In collaboration with Ceit and coordinated by Vicomtech, the consortium is made up of the University of the Basque Country (UPV), Automotive Intelligence Center (AIC), Basque Center for Applied Mathematics (BCAM), Tecnalia, Mondragon Unibertsitatea and Ikerlan.