面向车联网的车辆节点信誉评估方法  

A Reputation Evaluation Method for Vehicle Nodes in V2X

作　　者：田钊[1,2,3] 牛亚杰 佘维 刘炜[1,2,3] TIAN Zhao;NIU Yajie;SHE Wei;LIU Wei(School of Cyber Science and Engineering,Zhengzhou University,Zhengzhou 450002,China;Henan Key Laboratory of Network Cryptography Technology,Zhengzhou 450002,China;Zhengzhou Key Laboratory of Blockchain and Data Intelligence,Zhengzhou 450002,China;SongShan Laboratory,Zhengzhou 450002,China)

机构地区：[1]郑州大学网络空间安全学院,郑州450002 [2]河南省网络密码技术重点实验室,郑州450002 [3]郑州市区块链与数据智能重点实验室,郑州450002 [4]嵩山实验室,郑州450002

出　　处：《信息网络安全》2024年第5期719-731,共13页Netinfo Security

基　　金：河南省高等学校重点科研项目[24A520045];河南省网络密码技术重点实验室开放课题[LNCT2022-A04];嵩山实验室预研项目[YYYY022022003]。

摘　　要：车联网技术的发展促使了交通信息的交互与共享,能有效提升出行效率,但车联网的开放性使得交通实体容易受到恶意车辆的攻击,会造成严重的后果。针对上述问题,文章提出了一种面向车联网的车辆节点信誉评估方法。首先,提出了一种面向车联网的车辆节点分区区块链网络;然后,通过车辆节点之间的信任和基础设施对车辆的辅助信任来计算本地信誉值,借助深度学习算法动态计算全局信誉值,进而可基于全局信誉值确定最佳的数据共享节点;最后,对存储技术进行改进,利用分区区块链存储来保证信誉值、路况信息的不被篡改和可追溯性。仿真实验结果表明,本文所提的方法在通过信誉评估确定恶意节点的准确率和召回率都优于对比方法。The advancements in Vehicular Networks communication technologies facilitate the exchange and sharing of traffic information,thereby significantly enhancing travel efficiency.However,the openness of V2X networks increases the vulnerability of traffic entities to attacks from malicious vehicles,potentially leading to severe consequences.Addressing this issue,this paper proposed a reputation evaluation method for vehicle nodes in V2X.Initially,a partitioned blockchain network for vehicle nodes in Vehicular Networks was introduced.Subsequently,local reputation values were calculated based on trust among vehicle nodes and auxiliary trust from infrastructures,combined with the use of deep learning for dynamically computing global reputation values.This enabled the identification of optimal data sharing nodes based on global reputation scores.Finally,to enhance storage technology,partitioned blockchain technology was employed to ensure the integrity and traceability of reputation values and traffic information.Simulation results demonstrated that the proposed method outperformed comparative methods in accurately identifying malicious nodes,as evidenced by higher precision and recall rates.

关 键 词：车联网 信誉评估 区块链 深度学习 

分 类 号：TP309[自动化与计算机技术—计算机系统结构]