车联网中基于模糊评价密度聚类的共谋节点检测方法  

Collusion node detection method based on fuzzy evaluation density clustering in Internet of vehicles

作　　者：张海波[1,2] 王大斌 王汝言 王冬宇[3] ZHANG Haibo;WANG Dabin;WANG Ruyan;WANG Dongyu(School of Communication and Information Engineering,Chongqing University of Posts and Telecommunications,Chongqing 400065,China;Chongqing Key Laboratory of Ubiquitous Perception and Interconnection,Chongqing 400065,China;School of Artificial Intelligence,Beijing University of Posts and Telecommunications,Beijing 100876,China)

机构地区：[1]重庆邮电大学通信与信息工程学院,重庆400065 [2]泛在感知与互联重庆市重点实验室,重庆400065 [3]北京邮电大学人工智能学院,北京100876

出　　处：《通信学报》2023年第7期114-123,共10页Journal on Communications

基　　金：国家自然科学基金资助项目(No.61901071,No.61801065);长江学者和创新团队发展计划基金资助项目(No.IRT16R72);重庆市留创计划创新类基金资助项目(No.cx2020059)。

摘　　要：针对开放车联网环境中共谋恶意车辆带来的安全问题,设计了一种基于模糊评价密度聚类的共谋节点检测方法。首先,通过主观推荐信任、客观数据信任和历史信誉值实现车辆信誉更新。其次,设计两轮恶意车辆节点检测方法检测恶意车辆,第一轮利用模糊综合评价筛选出单个恶意车辆节点,第二轮根据单个恶意车辆节点,基于改进的密度聚类方法搜索出共谋恶意车辆节点,保证网络环境安全的可持续发展。实验结果表明,所提方法对恶意车辆有较高的识别率,当恶意车辆占比达到30%时,检测准确率仍能保持在90%以上。在不同恶意车辆占比下,检测召回率和检测F值整体保持较高数值,这表明所提方法具有较高的稳定性。针对总车辆节点的变化,性能评估指标的值变化幅度较小,仍然保持在80%~100%。In order to solve the security problems caused by collusion of malicious vehicles in the open Internet of vehicles environment,a collusion node detection method based on fuzzy evaluation density clustering was designed.The method realized vehicle reputation update through subjective recommendation trust,objective data trust and historical reputation.Two rounds malicious node detection method was designed to detect malicious vehicles,in the first round,fuzzy comprehensive evaluation was used to screen out a single malicious vehicle node,and in the second round,the conspiracy malicious vehicle nodes were searched by the improved density clustering method according to a single malicious vehicle node to ensure the sustainable development of network environment security.Experimental results show that the proposed method has a high recognition rate for malicious vehicles,and when the proportion of malicious vehicles reaches 30%,the detection accuracy can still remain above 90%.At the same time,under the proportion of different malicious vehicles,the detection recall rate and detection F value still remain higher value,indicating that the proposed method has a high stability.For the change of total vehicle nodes,the performance evaluation index shows a small fluctuation and remain within the range from 80% to 100%.

关 键 词：车联网 共谋攻击 模糊综合评价 密度聚类 

分 类 号：TP39[自动化与计算机技术—计算机应用技术]