Privacy-Preserving Trajectory Data Publishing by Dynamic Anonymization with Bounded Distortion

Publication of trajectory data that contain rich information vehicles in the dimensions time and space (location) enables online monitoring supervision motion offline traffic analysis for various management tasks. However, it also provides security holes privacy breaches as exposing individual’s to public may results attacks threatening safety. Therefore, increased attention has been made recently on protection publishing. existing methods, such generalization via anonymization suppression randomization, achieve by modifying original form a publishable trajectory, which significant distortion hence low utility. In this work, we propose privacy-preserving method called dynamic with bounded distortion. our method, individual trajectories set are mixed localized manner synthetic publishing, can protect location associated individuals ensure guaranteed utility published both individually collectively. Through experiments conducted real Guangzhou City Taxi statistics, evaluate performance proposed compare mainstream methods terms preservation against utilization. The show achieves better utilization than using globally static anonymization, without trading off attacks.