面向位置服务的车载自组织网络安全与隐私保护研究

Due to broadcasting nature of wireless channel and semi-trust of location servers, protecting privacy in the Location-Based Service (LBS) is very important in the Vehicular Ad Hoc Network (VANET) . Since VANET has several challenges in terms of frequent topology change, vehicles' movements limiting to roads, large scale of network, how to protect privacy of LBS is a very urgent problem if LBS can be used in the scale of VANET. Therefore, we proposed three schemes to protect LBS' privacy by employing Mix-zone, K-anonymity and game theory. Details are: 1) a scheme titled Mix-zone based game theory for pseudonym protection in VANET is given. In this scheme, firstly, a vehicle' s Mix-zone is dynamically formed when a vehicle requests it, and its size is determined on the prediction of location and the vehicle's level of privacy requirement. Secondly, we proposed a pseudonym changing game by adding rewards and penalties, in order to prevent a global attacker from guessing the mapping of old and new pseudonyms. 2) a scheme named continuous LBS location privacy based s-segment in VANET is proposed. The aim of this scheme is to improve the query accuracy and protect location privacy in the continuous LBS. Achieving it, we design an anonymous algorithm to get a anonymous cloaking region where at least s-segments and k-vehicles are locating, and use multi-object optimality to achieve the trade-off between query accuracy and location privacy. 3) Continuous LBS query content privacy protection based game theory scheme is proposed. We build up a Hide-Seek game from selfish vehicles and a rational attacker, and analyze their best strategies from complete information and incomplete information. This scheme can prevent a content collusion attack in the continuous LBS. The proposal will achieve the goal of publishing higher quality papers in the IEEE journals and top international conferences, and provide basic theories and techniques for spreading applications of LBS in VANET.

由于无线信道广播特性、位置服务器不可信任等特点，LBS业务给用户带来严重隐私威胁。因此，LBS隐私保护是车载自组织网络中LBS业务得以大规模推广的首要解决问题。本项目从车载自组织网络的拓扑变化快、车辆运动受道路限制等特点出发，提出1）基于博弈论的动态Mix-zone身份隐私保护算法，给出Mix-zone与不同车辆个性化隐私需求之间的关系，根据奖励与惩罚机制建立伪名更改博弈模型，阻止攻击者从新旧伪名推测出车辆的身份隐私；2）提出基于s-分路的连续查询位置隐私保护算法，达到匿名区域内的道路分布均衡目的，提高查询精确度，保护车辆的位置隐私；3）提出基于博弈论的连续查询内容隐私保护算法，从博弈论角度分析自私车辆和攻击者的最优策略，解决连续查询中所存在的内容联合攻击问题。本项目研究成果预期在IEEE期刊和高质量国际会议发表，为实现车载自组织网络中LBS业务的大规模推广提供前期理论基础和技术方案。

车载自组织网络(Vehicular Ad Hoc Network, VANET)被寄予厚望改善交通管理和提供安全位置服务的同时，也给人们带来极大的隐私威胁。因此，本报告围绕隐私问题开展相关内容研究，解决在位置服务器和RSUs半信任、车辆移动轨迹受限、拓扑动态变化等挑战下的隐私泄露威胁，从而保护用户隐私信息。具体工作如下：①提出了基于声望的身份隐私保护方法，通过声望模型促使更多的自私车辆参与伪名更新，从而显著提高车辆的个性化隐私，能够应用于车辆稀疏或密集的路网下的实时性业务。②提出了基于智能卡的用户认证方法，能够保护用户的匿名性，提供有效的用户认证协议，抵抗离线密码猜测攻击、智能卡丢失攻击、内部攻击等。③提出了连续位置服务的隐私保护方法，适用于多样化的路网环境下，能够抵抗连续查询攻击和最大速度攻击。④提出了基于簇匿名的位置服务隐私保护方法，通过路网和车辆信息转换，隐藏了路网信息和流量信息，从而阻止了攻击者通过路网信息和流量信息而推测用户的移动轨迹。⑤提出了基于博弈论的查询内容隐私保护方法，给出了基于Hide-and-Seek游戏的最优策略，从而更好的隐藏用户位置和查询内容之间的链接关系。⑥提出了基站位置隐私保护方法，通过引入虚假分组和虚假基站，避免了基站的位置隐私泄露问题，同时又能维持网络性能。上述的研究内容可为车载自组织网络的位置服务隐私保护提供前期理论探索和经验积累。相关理论和方法为实现车载自组织网络的位置服务业务的大规模推广和应用提供前期理论基础和技术方案。