Stealthy Data Corruption Attack Against Road Traffic Congestion Avoidance Applications

Abstract

Intelligent Transportation Systems (ITS) leverage open and real-time sharing of traffic data to enable more efficient transportation. However, the data exchanged over the vehicular network are easily corruptible via attacks known as misbehaviours. Misbehaviour detectors have been extensively developed but remain siloed and lack consideration of advanced attacks amalgamating multiple misbehaviours. These may be carried out as part of Advanced Persistent Threats. This paper presents a new approach to specifically designing stealthy data corruption attacks within ITS, and by extension in other data-reliant Cyber-Physical Systems. A Stackelberg security game is devised to model the actions of evasive attackers targeting congestion avoidance applications. The game is then solved to produce the optimal attack and defense strategies. The new stealthy attack achieves the intended long-term impact while improving evasion performance. This research direction exploring sophisticated attacks will allow to advance the design of robust misbehavior detection systems.