车载自组织网络安全关键技术研究

发布时间：2018-03-06 00:36

本文选题：车载自组织网络　切入点：安全　出处：《西安电子科技大学》2016年博士论文　论文类型：学位论文

【摘要】：车载自组织网络(vehicular ad hoc networks:VANET)是由车载节点、路边通信基础设施和后台服务器组成的自组织、异构的无线多跳网络,能够为驾驶员提供超视距的道路交通信息,实现碰撞避免、事故告警、路况感知等应用。由于VANET应用环境的特殊性,VANET攻击不仅影响网络正常功能,而且可能导致交通混乱甚至引发交通事故,因此提出完善、可靠的安全机制是VANET大规模部署和应用的重要前提。本文面向VANET安全这一主题,系统地分析了VANET攻击的类型及特点、VANET安全技术面临的困难和挑战,并围绕VANET安全中最突出和迫切的安全数据聚合、虚假消息检测等关键问题进行了积极的探索,获得了一定的理论与技术成果。针对VANET数据聚合过程中的安全威胁,提出了基于时-空关联特性的安全数据聚合方法。VANET数据聚合过程节省通信资源的同时也导致严重的安全隐患:数据提供者对原子数据的签名在聚合过程中被丢弃,消息不可否认性遭到破坏,给攻击者注入虚假数据提供了便利。现有的多种检测方法需要假设“诚实节点占多数”,难以抵御多节点合谋攻击。基于真实交通环境中车辆速度、密度等交通数据具有较强的空间关联性和时间自相关性的特点,提出一种基于时空关联的安全数据聚合机制。将数据聚合过程中发现异常数据的位置视为“不可观测区域”,基于人工神经网络方法利用其空间和时间上相邻的数据对其进行估值,根据估计值与车辆节点报告值之间的偏差程度评价原子数据的可信性,检测攻击者注入的虚假数据。为了降低检测过程导致的通信资源消耗,将攻击者-检测者之间的理性决策过程建模为两方不完全信息动态博弈,分析了博弈的混合策略精炼贝叶斯均衡解,提出了最大化检测者收益的优化检测策略。仿真结果显示本文方法的检测准确性和抗合谋攻击能力均显著优于现有的VARS和基于轨迹的检测方法,且能够实现检测准确性和资源消耗之间的折中平衡。针对内部敌手发起的虚假消息攻击,提出了安全多源数据融合的虚假告警消息检测方法。利用Dempster-Shafer证据理论提供的信度量化与合并方法融合车辆节点采集的多类型车载传感器数据,识别交通流模式,评估告警消息的可信性;为了扩大观测范围并增强检测结果的准确性,对多个观测车辆生成的基本信任分配(basic belief assignment:BBA)进行融合;为了排除观测车辆中的合谋敌手注入的虚假BBA,使用证据距离的概念量化两个BBA之间的相似性,并进一步使用离群点检测方法过滤虚假BBA。基于仿真方法评价了本文方法的性能,结果显示本文方法具有更强的抗合谋攻击能力,在不同的攻击者比例下检测率显著优于现有的基于阈值的检测方法。针对车辆节点自私性对VANET多节点合作的安全机制造成的影响,提出了联盟博弈的节点可信性评价方法。VANET中不端节点检测、安全路由等多种安全机制需要多个车辆节点彼此合作完成,自私节点的不合作行为能够对安全机制的性能造成不利影响。假设VANET车辆节点为理性的智能决策个体,使用合作博弈论建模车辆节点之间的合作行为,提出一种联盟博弈的节点信任关系建立方法:多个节点自发组成安全联盟,利用联盟成员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。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。针对V2I(vehicle to infrastructure)通信过程中车辆节点的隐私泄露风险,提出了基于网络编码的隐私保护V2I通信机制。V2I通信过程中窃听敌手通过数据流分析攻击可以确定数据包转发路径、追踪车辆节点移动轨迹并攻陷IEEE1609.2标准提出的多假名隐私保护机制。提出一种基于网络编码的隐私保护V2I安全通信机制,利用随机线性网络编码提供的消息分片、编码和缓存机制,消除多跳转发路径上的消息在大小、内容和发送时间上的关联,隐藏消息转发路径,保护车辆节点的身份隐私和位置隐私。安全性分析显示本文方法能够提供通信内容机密性、匿名性、不可关联性、不可追踪性等安全和隐私保护属性。仿真结果显示本文方法的计算负载显著低于现有的基于重加密技术的安全路由协议。
[Abstract]:Vehicular ad hoc network (vehicular ad hoc networks:VANET) is composed of vehicle nodes, self-organizing roadside communication infrastructure and backend server components, heterogeneous wireless multi hop networks, can provide traffic information for drivers over the horizon, avoiding collision, accident alarm, traffic sensing applications. Due to the special application environment of VANET the VANET attack not only affect the normal function of the network, and may lead to confusion and even lead to traffic accidents, therefore put forward to perfect, reliable safety mechanism is the important prerequisite for large-scale VANET deployment and application. The theme of this article for VANET security, systematically analyzes the types and characteristics of VANET attacks, VANET security technology facing difficulties and the challenge, and the most prominent around polymerization in VANET security and urgent safety data, explores the key issues of false news received detection. The theory and technology of results. According to the security threats of VANET data in the polymerization process, is proposed based on data security space correlation properties of.VANET polymerization polymerization process data save communication resources at the same time also lead to serious security risks: data provider for signature according to the atomic number are discarded during the polymerization process, the message can not think of the destruction, and provides convenience for the attacker to inject false data. Many existing detection methods need to assume the "honest majority node", it is difficult to resist the multi node collusion attacks. The speed of the vehicle in real traffic environment based on traffic density data has the characteristics of strong spatial correlation and time correlation, a polymerization mechanism safety data is proposed based on the spatial-temporal correlation. The data found in the polymerization process of the abnormal data position as "unobserved areas", based on Artificial Neural Networks By using the method of network space and time of adjacent data to estimate it, and the degree of deviation between the vehicle node evaluation report value of credibility of atomic data values according to the estimation, the false data detection. In order to reduce the attacker injects communication resources detection process leads to consumption, the rational decision-making process modeling between attackers - the detection the two party is not a complete information dynamic game analysis, mixed game theory refined Bias equilibrium solution, proposed to optimize the detection strategy to maximize detection yield. The simulation results show that detection accuracy of the proposed method and anti collusion performance are significantly better than the existing VARS and detection method based on trajectory, and can achieve the tradeoff between detection the accuracy and the consumption of resources. The false message attacks launched against internal adversaries, put forward to multi-source data fusion of false security Alarm message detection method. Multi sensor data fusion type vehicle vehicle acquisition node reliability quantification and merging method provided by Dempster-Shafer evidence theory, the traffic flow pattern recognition, credibility evaluation of alarm message; in order to expand the range and enhance the accuracy of observation results, the basic belief assignment to generate multiple observation vehicle (basic belief assignment:BBA) in order to eliminate false BBA fusion; adversary collusion in the vehicle into the observation, the similarity between the concept of quantization using evidence distance of two BBA, and further use of outlier detection method based on BBA. filter simulation false method to evaluate the performance of the method. The results show that this method has a better anti collusion attack ability in proportion the attacker under different detection rate was significantly better than the detection method based on threshold current. According to the vehicle nodes selfish Impact of the cooperation on the security mechanism of VANET multi node, the node.VANET credibility evaluation method of misconduct coalition game detection in a variety of security mechanisms, security routing requires multiple vehicle nodes to cooperate with each other, selfish nodes of non cooperative behavior can adversely affect the performance of the security mechanism. Assuming that VANET node for vehicle intelligent decision of individual rationality, using cooperative game theory, cooperative behavior modeling between vehicle nodes, presents a method to build the node trust relationship in a coalition game: a plurality of nodes spontaneously formed security alliance, cross location verification using wireless signal strength measurement of the members of the alliance, found that the malicious node position and inject forged false information. The income distribution function to meet the conditions of nuclear distribution encourage nodes to join the alliance; prove that this function satisfies individual rationality Conditions and effective conditions, the optimal strategy is to select the node cooperation, the formation of alliances with stability. The simulation results show that the effect of random deviation of the method can effectively reduce the wireless signal strength to verify the accuracy of the position, the false position detection rate of over 90%, and further strengthen the system for resistance to collusion attack. V2I (vehicle to infrastructure) vehicle node communication in the process of privacy disclosure risk, proposed the wiretap network privacy protection V2I encoding communication mechanism of.V2I communication process based on data flow analysis by adversary attacks can determine the packet forwarding path, multi kana privacy protection mechanism node mobile vehicle tracking trajectory and put forward the standard of IEEE1609.2. This fall a kind of network privacy protection encoding V2I security communication mechanism based on the random linear network encoding message slice, Encoding and caching mechanism to eliminate multi hop forwarding path of the message in the size of related content and time of sending the hidden message forwarding path, protect the vehicle node identity and location privacy. Security analysis shows that this method can provide the communication content confidentiality, anonymity, unlinkability, unlinkability security and privacy protection properties. Simulation results show that the computational load is significantly lower than that of the existing security routing protocol based on re encryption.

【学位授予单位】：西安电子科技大学
【学位级别】：博士
【学位授予年份】：2016
【分类号】：U495;TN915.08

【相似文献】