Detection of Data Injection Attacks on Predictive Vehicle Platooning

Abstract

In recent years, vehicle platooning has been proven effective in reducing fuel consumption and carbon emissions, as well as increasing road capacity. Subsequently, with the assistance of wireless communication, there has been a qualitative leap in platooning performance, but then some drawbacks also emerge. First of all, for the drawback of packet loss and delay, this article proposes a prediction-assisted platooning mechanism, in which each vehicle establishes its local platoon model to predict the motion of other vehicles, thereby reducing information latency. Then for the malicious data on communication, a detection algorithm based on several indicators is carried out. Finally, the advantages of the proposed platooning mechanism and detection algorithm are verified on a joint simulation platform that combines communication and traffic control. Furthermore, the verification results are more authentic and reliable due to the consideration of imperfections of realistic perception and interaction.