Complete bipartite anonymity: Confusing anonymous mobility traces for location privacy

Using mobile devices, people can easily obtain their location information, and access a wide range of location based services (LBSs). Many existing LBSs rely in accurate, continuous, and real-time streams of location information to provide quality of service guarantees. In this case, even if an user accesses LBSs anonymously, the identity of the user can still be revealed by analyzing the mobility trace. To protect user privacy, existing work sacrifice the quality of LBSs by degrading spatial and temporal accuracy. To achieve a better tradeoff between user privacy and the quality of service, we present a novel approach, Complete Bipartite Anonymity (CBA), to confuse the paths of nearby users by connecting different users' real traces with fake ones. CBA protects user privacy as users become indistinguishable after their paths are confused, the quality of service of LBSs is also guaranteed since users are able to report their accurate locations. We evaluate CBA by comparing the system and privacy performance with existing techniques such as Path Confusion or Query Obfuscation using a real-world data set, the results show that our scheme increases the chance for a user joining an anonymity group by 10 times in low user density areas, and reduces the resources consumed by about 90% for achieving the same anonymity degree.