车载自组网中联盟博弈的虚假数据检测策略
A False Data Detecting Scheme Based on Coalition Game for Vehicular Ad？？hoc Networks

针对车载自组网(vehicular ad？？hoc network, VANET)中多个恶意节点合谋的虚假数据攻击问题,提出了联盟博弈的虚假数据检测策略。该策略利用合作节点测量的无线信号强度进行交叉位置验证,发现伪造位置的恶意节点及其注入的虚假信息。为降低节点自私行为对检测准确性的影响,提出了满足“核”分配条件的收益分配函数激励节点加入联盟;证明了该函数满足个体理性条件和有效性条件,节点的最优策略是选择合作,形成的联盟具有稳定性。仿真结果表明:该策略能够有效激励节点合作,降低无线信号强度的随机偏差对位置验证准确性的影响,获得了超过90％的虚假位置检测率,并进一步增强了系统的抗合谋攻击能力,在不同的恶意节点比例下虚假告警消息检测率比基于表决的攻击者本地驱逐协议(local eviction of attackers by voting evaluators, LEAVE)提高了69.3％。
A false data detecting scheme based on coalition game is proposed to focus on collusive false data attack problem in vehicular ad？？hoc networks (VANET). The proposed scheme uses the measured radio signal strength at cooperative nodes to perform the position cross？？check and to detect the malicious nodes with false position and the false information injected by them. A payoff allocation function that satisfies the “core” allocation condition is proposed for stimulating nodes to join coalitions so that the impact of selfish behaviors of nodes on the accuracy of detecting can be reduced. It is proved that the proposed payoff allocation function satisfies an individual rational condition and an efficiency condition, and that the best strategy of nodes is to cooperate with each others and the formed coalition has the stability feature. Simulation results show that the proposed scheme can effectively stimulate nodes to cooperate, and reduce the impact of random deviation of radio signal strength on the accuracy of position verification. The detecting ratio of false positions is more than 90%. Furthermore, the ability for resisting collusive attacks is enhanced. A comparison with the LEAVE protocol show that the proposed scheme increases the detecting rate of false alert messages by 69.3% under various proportions of malicious nodes