Trajectory Analysis for Mobile Users with Cloaked Locations

Location based services (LBS) bring great convenience to mobile users. But, user privacy is threatened since a LBS server knows location information of all LBS requester, and it might leak location data in order to obtain extra profits. Some cloaked location methods were presented to ensure that user location cannot be obtained, but it might be inferred indirectly. In this paper, a trajectory analysis method is proposed to infer a real trajectory from a cloaked trajectory. First, prior knowledge base for geographical situation is constructed. Then, a hidden Markov model for trajectory analysis is established and trajectory recovery problem is reduced to the optimal state sequence prediction problem. Finally, trajectory analysis is fulfilled using the dynamic programming algorithm. Experiments show that there are still privacy disclosure risks even if a user adopts location cloaking approaches. In addition, our method provides a new way to evaluate the degree of privacy protection for LBS users. And it provides a reference to evaluate of privacy risks and improve privacy level from the view of privacy analysis.