Preserving Privacy in LBS against Attacks based on Concurrent Requests

This paper considers privacy threats involved in the use of location-based services (LBS) that have never been successfully addressed before. Previous work has mostly focused on anonymization techniques based on generalizing the user location contained in a given LBS request. We show that the presence of multiple concur- rent requests, and the distribution of different service parameters in the requests can significantly affect the level of privacy obtained by anonymization techniques. We provide a formal model of attacks in the above scenario, we propose defense techniques and we show their effectiveness by experimental evaluation.