Novel Intelligent BSM Falsification Attack Detection System Using Trusted Neighbor Vehicle Approach in IoV

Abstract

The proliferation of cyberattacks has emerged as a significant obstacle for advancing technologies such as the Internet of Things (IoT) and Internet of Vehicles (IoV) in recent times. Notably, cryptographic security measures have been implemented in IoV to counteract these cyberattacks. However, these security measures are inadequate when it comes to thwarting internal attackers within the network, as these attackers possess the necessary security credentials for authenticating basic safety messages (BSMs). The research community has made substantial contributions by proposing misbehavior detection systems (MDS) based on data-centric machine learning to identify and prevent internal attackers within IoV. Nevertheless, the existing MDSs in the literature rely on BSMs received from a single vehicle, thereby enabling internal attackers to manipulate their falsified BSMs and evade detection, resulting in a high incidence of false alarms. In this study, we introduce a new intelligent system for detecting falsified BSMs, employing a trusted neighbor vehicle approach (NIBFADS-UTVA)). Our approach demonstrates exceptional effectiveness, achieving an accuracy, precision, recall, and F1-Score all exceeding 99%.