A survey on misbehavior detection for connected and autonomous vehicles

• Published in: Vehicular Communications 2023-06, Vol.41, p.100586, Article 100586
• Main Authors: Bouchouia, Mohammed Lamine, Labiod, Houda, Jelassi, Ons, Monteuuis, Jean-Philippe, Jaballah, Wafa Ben, Petit, Jonathan, Zhang, Zonghua
• Format: Article
• Language: English
• Subjects: Connected and autonomous vehicles; Cybersecurity; Fault detection; Intelligent transportation systems; Machine learning; Misbehavior detection
• ISSN: 2214-2096
• DOI: 10.1016/j.vehcom.2023.100586

Connected and autonomous vehicles have recently emerged as promising technological solutions to optimize traffic congestion, prevent accidents, and enhance driving safety and efficiency. Since such vehicles are equipped with various embedded components connected through different communication technologies, their security has become a vital concern. Therefore, Misbehavior Detection plays a major role in enabling vehicles to quickly identify the security risks and adopt effective immediate countermeasures. In this paper, we present an in-depth study of misbehavior detection in connected and autonomous vehicles. We first develop a new definition of misbehavior based on a comprehensive analysis of the existing work related to both intentional (criminal) and unintentional misbehavior. The new definition lays a foundation to accommodate, characterize, and understand novel misbehaviors. We then extensively investigate the state-of-the-art solutions and provide a detailed taxonomy for a large family of machine learning algorithms used for misbehavior detection in the literature. We carefully review the available tools and datasets for misbehavior detection. Finally, we present open research directions and challenges.