Connected autonomous vehicles, or Vehicular Ad hoc Networks (VANETs), hold great promise, but concerns persist regarding safety, privacy, and security, particularly in the face of Sybil attacks, where malicious entities falsify neighboring traffic information. Despite advancements in detection techniques, many approaches suffer from processing delays and reliance on broad architecture, posing significant risks in mitigating attack damages. To address these concerns, our research proposes a Trust Aware Sybil Event Recognition (TASER) framework for assessing the integrity of vehicle data in VANETs. This framework evaluates information exchanged within local vehicle clusters, maintaining a cumulative trust metric for each vehicle based on reported data integrity. Suspicious entities failing to meet trust metric thresholds are statistically evaluated, and their legitimacy is challenged using directional antennas to verify their reported GPS locations. We evaluate our framework using the OMNeT++ discrete event simulator, SUMO traffic simulator, and VEINS interface with TraCI API. Our approach reduces attack detection times by up to 66% in urban scenarios, with accuracy varying by no more than 3% across simulations containing up to 30% Sybil nodes and operates without reliance on roadside infrastructure.

翻译：联网自动驾驶车辆，即车载自组织网络（VANETs），展现出巨大潜力，但其安全性、隐私性和安全性问题，尤其是在面对恶意实体伪造邻近交通信息的Sybil攻击时，仍令人担忧。尽管检测技术已取得进展，但许多方法存在处理延迟和对宏观架构依赖的问题，这在减轻攻击损害方面构成显著风险。为解决这些问题，本研究提出了一种信任感知Sybil事件识别（TASER）框架，用于评估VANETs中车辆数据的完整性。该框架评估本地车辆集群内交换的信息，基于上报的数据完整性为每辆车维护一个累积信任度量。对未能满足信任度量阈值的可疑实体进行统计评估，并使用定向天线验证其上报的GPS位置以质疑其合法性。我们使用OMNeT++离散事件模拟器、SUMO交通模拟器以及结合TraCI API的VEINS接口对框架进行评估。我们的方法在城市场景中将攻击检测时间减少了高达66%，在包含最多30% Sybil节点的模拟中准确率变化不超过3%，且无需依赖路边基础设施即可运行。