Dummy Generation Based on User-Movement Estimation for Location Privacy Protection

Abstract

Location-based services (LBSs) have been becoming more common due to the prevalence of GPS-enabled devices. While LBSs bring many benefits for our daily lives, location information may reveal private information, rendering an important problem of protecting location privacy of users. To anonymize the locations of users, we focus on dummy-based approaches that generate dummies and their locations are sent along with the actual location of a user to an LBS provider. Although several existing studies developed dummy-based techniques, they assume unrealistic user mobility, e.g., users keep moving and do not stop or follow a pre-defined movement plan precisely. In this paper, we remove the unrealistic assumptions and require much easier input with respect to user movement, i.e., only a set of visiting points. Under the assumption, we propose a dummy generation method, estimation-based dummy trajectory generation (Edge). Based on the given visiting points, Edge estimates a user-movement plan and designs trajectories of dummies so that the adversaries cannot distinguish the user from dummies. We conduct extensive experiments using real map information, and the results show the efficiency and effectiveness of Edge.