Hiding stars with fireworks: location privacy through camouflage

Individuals face privacy risks when providing personal location data to potentially untrusted location based services (LBSs). We develop and demonstrate CacheCloak, a system that enables realtime anonymization of location data. In CacheCloak, a trusted anonymizing server generates mobility predictions from historical data and submits intersecting predicted paths simultaneously to the LBS. Each new predicted path is made to intersect with other users' paths, ensuring that no individual user's path can be reliably tracked over time. Mobile users retrieve cached query responses for successive new locations from the trusted server, triggering new prediction only when no cached response is available for their current locations. A simulated hostile LBS with detailed mobility pattern data attempts to track users of CacheCloak, generating a quantitative measure of location privacy over time. GPS data from a GIS-based traffic simulation in an urban environment shows that CacheCloak can achieve realtime location privacy without loss of location accuracy or availability.