基于轮廓泛化的位置隐私保护模型及方法

移动用户在连续位置服务过程中会产生大量的相关位置,攻击者可通过用户轮廓将其关联成位置轨迹。由于位置轨迹含有更多时空相关信息,使得攻击者更易获取用户个人隐私。针对这种情况,通过二分图刻画描述了轮廓和位置之间的关联关系,并基于该关联关系建立了Lθk隐私保护模型以及可抵抗轮廓关联攻击的假位置生成方法。该方法为每个服务位置生成轮廓信息相似的虚假位置,模糊了轮廓与真实位置之间的关联关系,保护了用户在导航或最近邻查询过程中的位置隐私。最后,性能分析及实验验证进一步证实,所提方法可提供较高的隐私保护级别和较好的算法执行效率。

Location privacy protection model and algorithm based on profiles generalization

When enjoying the continuous location based services (LBSs), a mobile user will generate a series of related locations, and the adversary can utilize user’s profiles to correlate these locations and assemble them into a trajectory. As the trajectory contains much more temporal and spatial information than discrete locations, which makes the adversary easy to infer user’s privacy. In order to cope with the privacy leaks, a binary chart is used to depict the relationship between user’s profiles and related locations. By the relationship depicted by the binary chart, an Lθk privacy preservation model and an inference resistant dummy algorithm (IRDA) are proposed. The IRDA can generate several dummy locations for each service location with a similar profile, and obfuscate the relationship between the profile and real location. Therefore, it can preserve the location privacy of users in the service of navigation or nearest neighbor query, and achieve the continuous location privacy preservation. Finally, the property analysis and experiment results further verify that the proposed algorithm can provide a higher location privacy level and have better algorithm performance efficiency.