Scalable Collaborative Intrusion Detection in Autonomous Vehicular Networks: A hierarchical framework based on game theory

Abstract

Autonomous Vehicles (AVs) networks are vulnerable to security attacks targeting data integrity and service availability. To successfully operate in emerging automated highway systems (AHS), securing communications among the AVs and with the transportation infrastructure is paramount. An Intrusion Detection System (IDS) is the foremost defense artifact against malicious AV activities. However, individual IDS on board of the AV may not be sufficient enough against diverse and increasingly emerging new threats. In addition, not all security vendors synchronize their update of new IDS security features. Therefore, the collaboration among individual IDSs of AVs enables the vehicles to benefit from the collective knowledge and information of their neighbors and achieve more accurate intrusion detection. This paper presents a hierarchical framework for a Collaborative Intrusion Detection Network (CIDN) that allows the AV platoons to share detection knowledge based on trust and increase the accuracy of their individual IDS. The proposed CIDN involves intra-platoon and inter-platoon collaborations to improve scalability. The experimental results show that collaborative IDSs yield higher accuracy than individual IDSs in detecting misbehaviors in AV networks while achieving scalability.