Cover locations: availing location-based services without revealing the location

  • Authors:

  • Sai Teja Peddinti;Avis Dsouza;Nitesh Saxena

  • Affiliations:

  • Polytechnic Institute of New York University, Brooklyn, NY, USA;Polytechnic Institute of New York University, Brooklyn, NY, USA;University of Alabama, Birmingham, Birmingham, AL, USA

  • Venue:
  • Proceedings of the 10th annual ACM workshop on Privacy in the electronic society
  • Year:
  • 2011

Quantified Score

Hi-index 0.00

Visualization

Abstract

Location-Based Services (LBSs) have been gaining popularity due to a wide range of interesting and important applications being developed. However, the users availing such services are concerned about their location privacy, in that they are forced to reveal their sensitive location information to untrusted third-parties. In this paper, we propose a new privacy-preserving approach, Cover Locations, which allows a user to access an LBS without revealing his/her actual location. Based on its current location, the user's device queries for a few specifically chosen surrounding locations and constructs the results corresponding to its location from the results obtained for each queried location. Since the user location does not leave the user's device - as either a latitude and longitude pair, or as an obfuscated region - the user is guaranteed very high level of privacy. The Cover Locations approach only requires minimal changes on the user's device and can be readily deployed by privacy-conscious users. An adversary, trying to identify the user location, can only resolve the location to few triangular regions and not to the actual location itself. We evaluate the privacy provided by Cover Locations based on the number of locations queried and the total area under the resolved triangular regions. We also ascertain the robustness of Cover Locations approach when the adversary has access to a short-term user history, employing machine learning techniques. Overall, our results show that the proposed solution, which requires minor computations without the need for any out-of-band information such as traffic densities in a region or the road network information, is superior to other client-based solutions.