Real-Time Privacy-Preserving Data Release Over Vehicle Trajectory

Intelligent connected vehicle trajectory data are of great value for data mining applications such as traffic management and commercial institutions. However, the leakage of sensitive trajectory makes the user hesitate to use the system if no privacy-preserving mechanism is adopted. In this paper, we propose a privacy-preserving mechanism with differential privacy called RPTR, which protects a vehicle's real-time trajectory data release. First, RPTR adopts a dynamic sampling method to process the trajectory data to meet the application load and practicability. Meanwhile, to ensure the data availability, ensemble Kalman filter based on users’ position transfer probability matrix is used in the prediction calculation. Also, we construct the privacy budget allocation method based on regional privacy weight to provide better protection for regions with high user density. Through our analysis and experiments, RPTR not only protects the privacy of real-time trajectory data but also guarantees the data availability.