基于统计差分的轨迹隐私保护方法的研究

发布时间：2018-10-18 10:06

【摘要】：伴随着信息社会和定位技术的快速发展,基于位置的服务已经成为生活中不可缺少的一部分,并使人们生活变得更加便捷。与此同时,人们也将面临着轨迹隐私泄露的问题。由于人们对轨迹隐私的关注度不断地提升,使得轨迹隐私保护问题成为当前研究的热点。而传统的轨迹隐私保护技术存在部分局限性和缺陷,为此,2006年Dwork提出差分隐私保护技术,迅速地获得国内外学者的关注,并将其应用于轨迹隐私保护领域。基于差分隐私保护的轨迹数据发布方法可以有效地保护轨迹隐私。但是,现有的差分隐私保护方法仍存在若干问题亟待解决:一方面,现有的差分隐私保护方法中采用添加随机噪音机制实现差分隐私保护,存在添加噪音量不可控的问题。另一方面,传统的轨迹隐私保护方法缺少相应的后置处理技术。因此,添加适量噪音和设计后置处理技术是提高发布轨迹数据可用性的关键因素,这也就是本文需要研究解决的问题。为解决现有差分隐私保护方法中存有的添加随机噪音和相关后置处理技术的问题,本文提出一种新型的基于统计差分的轨迹隐私保护方法。首先,该方法根据轨迹所具有的马尔科夫过程特性计算相应位置节点的敏感度;然后,利用位置节点的敏感度和相应的敏感度阈值添加适量的拉普拉斯噪音,并构建噪音前缀树;最后,设计和实现相应的后置处理技术——一致性约束,对噪音前缀树进行一致性约束处理,优化噪音前缀树,从而提高发布轨迹数据的可用性。实验证实,基于统计差分的轨迹隐私保护方法既实现了差分隐私保护,又有效地提高了发布轨迹数的可用性。
[Abstract]:With the rapid development of information society and positioning technology, location-based services have become an indispensable part of life, and make people's lives more convenient. At the same time, people will also be faced with the issue of trajectory privacy leakage. Due to the increasing attention to trajectory privacy, the protection of trajectory privacy has become a hot topic. However, the traditional trajectory privacy protection technology has some limitations and defects. For this reason, Dwork proposed differential privacy protection technology in 2006, which quickly attracted the attention of scholars at home and abroad, and applied it to the field of trajectory privacy protection. Trajectory data distribution method based on differential privacy protection can effectively protect trajectory privacy. However, there are still some problems to be solved in the existing differential privacy protection methods. On the one hand, in the existing differential privacy protection methods, adding random noise mechanism is used to realize differential privacy protection, and the problem of adding noise is not controllable. On the other hand, the traditional path privacy protection method lacks the corresponding post-processing technology. Therefore, adding appropriate amount of noise and designing post-processing technology are the key factors to improve the usability of release trajectory data, which is the problem that needs to be solved in this paper. In order to solve the problem of adding random noise and related post-processing techniques in existing differential privacy protection methods, a new path privacy protection method based on statistical difference is proposed in this paper. Firstly, the sensitivity of the corresponding position node is calculated according to the Markov process characteristic of the locus, and then the appropriate Laplace noise is added by using the sensitivity of the position node and the corresponding sensitivity threshold. The noise prefix tree is constructed. Finally, the corresponding post-processing technique, consistency constraint, is designed and implemented, and the noise prefix tree is optimized to improve the usability of the published track data. The experimental results show that the path privacy protection method based on statistical difference not only realizes differential privacy protection, but also effectively improves the availability of the number of published trajectories.
【学位授予单位】：郑州大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP309

【相似文献】