Leveraging Geographic Information and Social Indicators for Misbehavior Detection in VANETs

With the development of Vehicular Ad hoc Networks (VANETs), the security threats to the communication of VANETs are also increasing. Misbehaviors including fake messages attacks, characterized by messages containing misleading traffic information will significantly interfere with the cooperation between vehicles, and reduce the message transmission and service capabilities of VANETs. Therefore, it is necessary to propose an effective misbehavior detection scheme to ensure the reliable communication of VANETs. In addition to considering the dynamic network topology of VANETs, the social relationship between vehicles can also be applied as a relatively stable indicator to evaluate vehicles. This paper proposes a misbehavior detection scheme based on geographic information in vehicular social networks, to address potential security threats in VANETs. The proposed scheme constructs social relationships by analyzing similarities in geographic movement patterns of vehicles, and also evaluates the message trust from time and spatial factors. Finally, the proposed scheme integrates message trust with the established social relationships, enhancing the overall misbehavior detection accuracy, and then utilizes a reputation system to dynamically adjust the involvement of vehicles in misbehavior detection based on their historical behavior. Moreover, simulation results extensively demonstrate that our proposed scheme effectively detects malicious vehicles, performing consistently well across different ratios of such vehicles and movement patterns.