关联概率不可区分的位置隐私保护方法

Correlation probability indistinguishable location privacy protection algorithm

首先量化了在快照查询服务和连续查询服务中攻击者可能通过关联关系建立的关联概率攻击方法。然后,针对这些攻击方法提出了与之对应的基于广义差分隐私的隐私保护模型。基于建立的隐私保护模型设计了基于位置偏移产生关联概率不可区分的隐私保护方法,并证明了这种方法的隐私保护效力。最后,通过实验进一步验证所提模型和方法的隐私保护效力和算法执行效率。

The attack of correlation probability used by the adversary in snapshot query and continuous query services were measured. Then, on account of these attacks, a privacy protection framework based on the differential privacy and a location-shift scheme to achieve the indistinguishable of correlation probabilities was provided, and the protection effec-tiveness of this method was proved. At last, security analysis and the experiment results further verify the protection ef-fectiveness and guarantee the execution efficiency of the proposed scheme.