Trajectory privacy protection scheme based on differential privacy

Abstract

To solve the problem that the current sampling mechanism and data obfuscation method may raise insufficient data availability and privacy protection, a trajectory privacy protection scheme based on differential privacy was proposed.A new efficient sampling model based on time generalization and spatial segmentation was presented, and a k-means clustering algorithm was designed to process sampling data.By employing the differential privacy mechanism, the trajectory data was disturbed to solve the user privacy leaking problem caused by the attacker with powerful background knowledge.Simultaneously, to respond to the error boundary of the query range of pandemic, an effective prediction mechanism was designed to ensure the availability of released public track data.Simulation results demonstrate that compared with the existing trajectory differential privacy protection methods, the proposed scheme has obvious advantages in terms of processing efficiency, privacy protection intensity, and data availability.

Keywords

• differential privacy;trajectory privacy;data sampling;exponential mechanism;data publishing