基于个性化隐私需求的查询隐私保护算法研究

Research of query privacy protection algorithm with personalized requirements of privacy

现有大多数基于位置服务(location based service,LBS)的隐私保护算法都将对用户位置隐私的保护等同于对整个LBS查询服务隐私的保护.但是,在用户位置信息已知的前提下,这些算法有可能面临推断攻击.在考虑用户个性化隐私需求的情况下,基于四分树结构提出了能够避免此类推断攻击的隐私保护算法;为了有效的减小隐惹区域的大小基于半象限的定义对该算法进行了进一步优化.最后,通过仿真实验验证了算法抵御推理攻击的有效性.

Location privacy protection and query privacy protecton is considered to be equivalent in most existing LBS （ location - based service） privacy protection algorithms. However, under the prem- ise of the user location information is known, these algorithms will lead to inferring attack. In this pa- per, we proposed an algorithm based on quadtree to avoid this kind of attack, and to satisfy the user＇ s personalized query privacy. Furthermore, in order to rerude the size of cloaking area, we optimized the algorithm based on half quadrant defination. Finally, the effectiveness of the algorithms to resist inferring attack is verified by simulation experiment.