In this paper, we study potential inference attacks targeting location-based service (LBS) users. In particular, we introduce a new model for privacy protection, provides heuristic defence techniques to protect users’ privacy from such attacks, and present the results of experiments performed to evaluate the heuristics. Potential attackers who gain access to supplemental information may infer sensitive information such as location, identity or lifestyle about a user querying an LBS. Supplemental information used includes the times when queries are submitted, speed limits, and travel times for the underlying road network, or residential/commercial address directories. Our objective here is to prevent attackers from connecting external information to user queries. To address this objective, we introduce the notion of (i, j)-privacy. The novel (i, j)-privacy model generalises previous privacy models and allows users to customise their own privacy levels. To implement (i, j)-privacy, we have designed several heuristics. Although these are heuristic approaches, they do provide exact responses for user queries. We evaluate these heuristics experimentally on different road networks. We study the impact of a number of input parameters (mainly geometric) and present the results here. Our experiments demonstrate that, for realistic user settings, our algorithms provide results rapidly and of high quality.

Additional Metadata
Keywords inference attacks, k–anonymity, location and query privacy, Location-based services, obfuscation
Persistent URL
Journal Journal of Location Based Services
Nussbaum, D, Omran, M.T. (Masoud T.), & Sack, J.-R. (Jörg-Rüdiger). (2017). Maintaining anonymity using -privacy. Journal of Location Based Services, 1–28. doi:10.1080/17489725.2017.1363419