车联网环境下基于重复博弈的恶意车辆节点检测机制

针对车联网内部存在的虚假信息攻击，以及节点动态变化快及密集程度不同造成的恶意车辆节点检测机制效率低下，提出了一种基于重复博弈的恶意车辆节点检测机制。首先，根据车辆在信息交互中的行为建立重复博弈模型，并利用生成的节点收益计算出信任值与动态阈值，经二者比较，筛选出可疑的恶意车辆节点；其次，通过权值投票算法从可疑的恶意车辆节点中判定出恶意车辆节点；最后，从邻居列表中选取信任值最高的下一跳车辆节点进行合作。仿真和分析表明，与现有的相关机制相比，该机制提高了对虚假信息攻击的检测率，降低了误检率。

Malicious vehicle node detection mechanism based on repeated game in VANET

In view of Internet of vehicles within the false information attack, and inefficiency of malicious vehicle nodes detection mechanism caused by fast node dynamic change and intensive different, this paper proposed a malicious nodes vehicle detection mechanism based on repeated game. Firstly, according to the behaviors of the vehicle in the information interaction it established the repeated game mode, and used the generated node income to calculate the trust value and the dynamic threshold. By comparison, it screened out the suspicious malicious vehicle nodes. Secondly, it identified the malicious nodes vehicles by weights of voting algorithm from suspected malicious nodes vehicles. Finally, it selected the next-hop vehicle node with the highest trust value from the neighbor list to cooperate. Simulation and analysis show that compared with the existing mechanism, this mechanism improves the detection rate of false information attack, reduces the error detection rate.