位置服务中基于博弈分析的轨迹隐私保护机制

发布时间：2018-03-22 16:10

本文选题：位置服务　切入点：位置隐私　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】：位置服务提供丰富的应用服务,逐渐渗入到人们的日常生活,如自然环境、公共设施、社交活动等位置相关的感知服务,叫车、导航、广告推送等基于位置的请求服务,以及提供轨迹信息给科研人员和企业分析与挖掘的位置信息发布服务。位置服务在带来服务便利的同时,也泄露了移动用户的位置信息。服务器在收集感知数据,提供位置查询服务及轨迹发布服务时可直接或间接获取用户的位置,攻击者也可能攻击服务器获取用户的位置信息,而且攻击者还可能从用户之间的交互、发布的位置轨迹数据中推断某用户的位置。本文围绕位置信息感知、使用和发布等三类位置服务中的隐私保护问题,在系统分析现有位置隐私和轨迹隐私保护技术基础上,利用博弈论分析并设计相应的隐私保护机制。具体研究内容如下：(1)针对位置信息感知服务,研究群智感知位置服务中的隐私保护机制。在群智感知位置服务中,感知准确性与用户隐私是一对矛盾体。服务器收集的自然环境、公共基础设施和社交活动信息越多越丰富,服务器获取的感知信息越准确；然而,用户上传的感知信息越多越丰富,也就越容易泄露用户的位置隐私。针对这一挑战,本文提出了一种隐私保护的位置感知数据上传机制。通过不完全信息博弈构建用户上传行为博弈,每个终端用户根据服务器提供的感知服务质量反馈信息,权衡群智感知服务质量和自身的位置隐私,独立地决定是否上传。根据用户上传博弈的纳什均衡,本文分析终端用户上传行为与群智感知服务质量和位置隐私泄露之间的关系,在此基础上设计合理的用户终端上传策略,在满足基本的群智感知服务质量需求的前提下,最大化用户的隐私保护度。通过真实轨迹数据的仿真实验,证实该机制能够兼顾群智感知服务质量和用户隐私,指导终端用户选择了最优的上传策略,最大化了用户的效用。(2)针对基于位置的请求服务,研究LBS查询服务中的隐私保护机制。现有的LBS位置隐私保护机制分为集中式和分布式方法。集中式方法依赖可信匿名中心匿名或混淆用户的查询请求,但可信的匿名中心通常不存在,而且用户也不愿信任匿名中心。分布式方法让用户自己产生假查询来保护用户的位置和查询隐私,但假查询会增加用户的存储、通信和计算开销,而且假查询很容易被某些网络追踪手段排除,该方法很难以较低的开销达到较好的匿名保护度。因此,本文提出了一种基于用户协作的隐私保护查询方案,通过与附近用户相互协作形成匿名组,共同制定匿名组内用户的查询策略。该方案利用贝叶斯博弈分析匿名组用户选择不同查询策略对k-匿名成功率的影响,基于博弈的纳什均衡,设定合理的博弈参数,促进用户相互协作,提高k-匿名成功率,最大化用户的效用。通过博弈纳什均衡结果分析和仿真实验,证实了该查询策略生成算法能够在用户数小于k和大于等于k的情形下,指导组内用户产生了最优的查询策略,保证匿名组的k-匿名成功率。(3)针对位置发布服务,研究轨迹大数据发布服务中的隐私保护机制。虽然研究人员提出了很多轨迹隐私保护方法,如添加假轨迹,减少轨迹中的抽样点,对抽样点添加噪声等,但这些方法大多针对特定的轨迹分析场景。不同场景对轨迹数据的真实性需求不一样,这也造成了轨迹隐私保护机制在不同场景中的效果差异,而且隐私保护效果还受不同攻击策略的影响。针对这一问题,本文提出了一种隐私保护策略优化选择算法,指导第三方数据平台选择最优的隐私保护策略。首先采用完全信息博弈和不完全信息博弈建模攻击者和防御者之间对抗行为,通过攻防博弈分析在不同的攻击策略下防御者采用不同隐私保护策略达到的隐私保护效果；然后基于该分析结论指导第三方数据平台根据不同的数据真实性需求选择隐私保护策略,使得隐私保护策略在攻击者采用最优攻击策略时仍能展现较好的性能,最大化防御者的效用。通过博弈纳什均衡和真实轨迹数据的仿真实验表明,隐私保护策略优化选择算法指导第三方数据平台选择的隐私保护策略优于其他隐私保护策略。
[Abstract]:Location services provide a rich application services, gradually infiltrated into people's daily life, such as natural environment, public facilities, social activities related to location aware service, called the car navigation, advertising push service request based on the location, and provide information to track position information of scientific research personnel and business analysis and mining publishing service location service in service. Bring convenience, but also reveal the location information of a mobile user. The server data collected in perception, position can be direct or indirect access to the user's position and trajectory query service publishing service, the attacker may attack the server to get the user location information, and the attackers may also from the interaction between users. That a user's position trajectory data released. This paper focuses on the position of information privacy protection awareness, use and release of three position in service The problem, in the system analysis of the existing location privacy and trajectory privacy protection technology on the basis of using game theory to analyze and design the corresponding privacy protection mechanism. The specific contents are as follows: (1) according to the position information service, privacy protection mechanism of crowdsourcing location service research. In the crowdsourcing location service, perceived accuracy with the user privacy is a contradiction. The server collects natural environment, public infrastructure and social activities more abundant information, the server obtains more accurate sensing information; however, the perception of information more users to upload more abundant, but also easier to reveal the user's location privacy. In response to this challenge, is presented in this paper. Location aware data upload a mechanism of privacy protection. Through the incomplete information game user upload behavior game, each user terminal server provides service according to perception Service quality feedback information, weigh the location privacy of swarm intelligent perception of service quality and their own, independently decide whether to upload. According to the Nash equilibrium game users upload, this paper analyzes the relationship between the terminal users to upload behavior and swarm intelligent perception of service quality and location privacy, users upload strategies on the basis of reasonable design, in the premise of to meet the basic needs of swarm intelligence perceived service quality under the maximum user privacy protection. Through the simulation experiment, real trajectory data, confirmed that the mechanism of swarm intelligence can take into account the perceived service quality and user privacy, guide the end user to select the optimal upload strategy, to maximize the utility of users. (2) according to the request service location based on LBS queries on privacy protection services. Existing LBS location privacy protection mechanism is divided into centralized and distributed centralized method method. Rely on trusted anonymous anonymous center or confuse the user's request, but the anonymous center credible usually does not exist, but the user has to trust the anonymous center. Distributed method allows you to produce false query to protect the user's location and query privacy, but the false query will increase the user's storage, communication and computation overhead, and the query is false easy to be some network tracking means excluded, this method is difficult to lower overhead to achieve better anonymity degree. Therefore, this paper proposes a query scheme for privacy protection based on user cooperation, through mutual cooperation with nearby users to form the anonymous group, jointly developed the user anonymous group within the query strategy. The scheme analysis of anonymous group the user selects a different query strategy of k- anonymous success rate by using the Bias game, the Nash equilibrium of the game based on the game, set reasonable parameters, promote users Mutual cooperation, improve the k- anonymous success rate, to maximize the utility of users. Through the game Nash equilibrium analysis and simulation proved that the algorithm can generate the query strategy is smaller than K and greater than or equal to K in the number of users under the guidance of users within a group have the optimal search strategy to ensure the success of the anonymous anonymous group k- rate. (3) according to the location of distribution services, privacy protection services research track big data release. Although researchers have proposed many methods such as adding trajectory privacy, false track, reduce the sampling points in the track, on the point of adding noise, but most of these methods for specific trajectory analysis real scene. The demand of different scenarios on trajectory data is not the same, this is caused by the difference in effect of trajectory privacy protection mechanism in different scenarios, and the privacy protection effect is also affected by different attack strategies. To solve this problem, this paper proposes a privacy protection strategy optimization algorithm, guiding the third party platform data privacy protection strategy. Firstly, between the optimal game of complete information and incomplete information game modeling of attackers and defense against the anti attack behavior, game analysis of privacy protection effect in defense of different attack strategies using different strategies to achieve privacy protection; then the analysis conclusion guiding the third party data platform according to the real demand choice of privacy protection strategy based on different data, the privacy protection strategy in an attacker using the optimal attack strategy still can show better performance and maximize the utility of defense. Through the simulation experiment and the real trajectory of Nash equilibrium game the data show that privacy protection strategy selection and optimization of privacy protection strategy guiding the third party data platform selection Better than other privacy protection strategies.

【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】：TP309

【相似文献】