Privacy Preservation Using Multiple Mix Zones

Abstract

In this chapter, we investigate the optimal multiple mix zones placement problem for location privacy preservation. We model the area covered by location-based services as a graph, where all vertices (POIs) are considered as candidates for mix zone deployment. In order to protect mobile users from side information based inferential attacks, we propose to use pairwise vertex association to characterize the linkability of the POIs along a user’s trajectory on the map. To achieve maximum privacy preservation, we formulate the optimization problem with the objective of maximizing the overall discontinuity of all possible trajectories on the road network and subject to deployment cost, traffic density, and differentiated privacy priority constraints. For each road segment and intersection, the traffic density effect in terms of entropy is also taken into account. We design three heuristic algorithms corresponding to different traffic scenarios and privacy preservation levels as practical and efficient solutions to the NP-hard optimization problem. Through extensive simulations based on realistic mobile user data traces, we demonstrate that our solution yields satisfactory performance in reducing the success rate of inferential attacks. The mathematical modeling and performance results presented in this chapter offer both theoretical and practical guidance to multiple mix zones placement in mobile networks for protecting users’ location privacy.