Security is crucial for cyber-physical systems, such as a network of Connected and Automated Vehicles (CAVs) cooperating to navigate through a road network safely. In this paper, we tackle the security of a cooperating network of CAVs in conflict areas by identifying the critical adversarial objectives from the point of view of uncooperative/malicious agents from our preliminary study, which are (i) safety violations resulting in collisions, and (ii) traffic jams. We utilize a trust framework (and our work doesn't depend on the specific choice of trust/reputation framework) to propose a resilient control and coordination framework that mitigates the effects of such agents and guarantees safe coordination. A class of attacks that can be used to achieve the adversarial objectives is Sybil attacks, which we use to validate our proposed framework through simulation studies. Besides that, we propose an attack detection and mitigation scheme using the trust framework. The simulation results demonstrate that our proposed scheme can detect fake CAVs during a Sybil attack, guarantee safe coordination, and mitigate their effects.

翻译：安全性对于赛博物理系统至关重要，例如一队网联及自动驾驶车辆（CAVs）协同安全地穿行于道路网络。本文通过前期研究从非合作/恶意代理的视角识别关键敌对目标，即（i）导致碰撞的安全违规行为，以及（ii）交通拥堵，从而解决冲突区域内CAV协作网络的安全问题。我们利用信任框架（且本研究不依赖于特定的信任/声誉框架选择）提出一种弹性控制与协同框架，以减轻此类代理的影响并保障安全协同。实现敌对目标的一类攻击是女巫攻击，我们通过仿真研究验证所提框架的有效性。此外，我们提出一种基于信任框架的攻击检测与缓解方案。仿真结果表明，本方案能在女巫攻击中检测虚假CAV、保障安全协同并减轻其影响。