Nearby friend discovery with Geo-indistinguishability to stalkers

Nearby friend discovery is a popular location based service (LBS), which allows you to discover nearby like-minded people, and make friends with them. The main privacy threat of this service is that the disclosure of locations leaves opportunities for stalkers. Although location privacy preservation in LBS has recently received much attention, few works have been done on privacy-aware nearby friend discovery, where people want to discover nearby friends without exposing their private locations to arbitrary strangers. Unlike most of other LBS services, nearby friend discovery needs to consider the privacy of both of the two communicating entities, i.e., the user who is searching for nearby people, and the user who is being discovered by others. This unique property makes it a great challenge to protect users' location privacy while providing satisfactory service quality. This paper presents the first research addressing this issue by combining the location approximation technique and the homomorphic cryptography. We show that the proposed scheme provides formal privacy guarantees for the LBS users, and still achieves satisfactory quality of the LBS.