A ε-sensitive indistinguishable scheme for privacy preserving

In general, continuous queries in location-based service usually point to a predetermined destination, as the destination usually means the most sensitive point for a user, the process of moving will show a trend of sensitivity increasing. As a result, this trend can be used to infer the destination as well as some other personal privacy such as workspace and home addresses. In order to cope with this privacy issue, in this paper we first formalize the pattern of sensitivity increasing and study this pattern to discuss the effectiveness on inferring the security that disposed by several currently used privacy preservation schemes. Then, we propose a ε-sensitive indistinguishable scheme to address the attack of sensitivity inference. Specifically, the proposed scheme utilizes a grid of Voronoi diagram to depict the sensitive contours and adds dummies according to the conception of differential privacy to cloak the trend of sensitivity increasing. At last, we illustrate the security analysis to verify the capacity of privacy preservation and we also verify the performance of this scheme with experimental verification in both Euclidean space and road networks and compare it with other schemes.