Publié le 28/11/2024
In this thesis, we focus on misbehavior detection in cooperative intelligent transport systems, with particular contributions in the area of cooperative perception. Cooperative perception technologies emphasize interactions between multiple agents and fuse shared data from different sources to overcome the limitations of single-agent perception. However, the additional exchanged data introduces more challenges in ensuring its integrity and authenticity, particularly against semantic-based attacks and failures. Our work addresses two key components : the implementation of the ExFMD-CP framework and the evidence-based misbehavior detection approach. We apply Belief Theory and Subjective Logic, which enable the assessment of uncertainty, ambiguity, imprecision, and the combination of multiple data sources to detect malicious agents. The use of various theoretical tools and fusion operators has significantly improved the performance of the proposed approach.
La soutenance aura lieu le 13 décembre à 10h à l’IRT SystemX à l’Inria (salle Robin Milner, 48 Rue Barrault, 75013 Paris).