DP-ARST: Adaptability Regional Trajectory Synthesis with Differential Privacy

The use of vehicle trajectory data has become increasingly widespread, raising concerns about the privacy and security of sensitive information contained within this data. Synthetic trajectory with differential privacy has emerged as the primary data protection method. However, traditional approaches often lead to over-protection and under-protection, particularly in complex geographical environments. To address these challenges, we propose the adaptive regional trajectory synthesis (DP-ARST) method. DP-ARST refines trajectory information, enabling detailed analysis while preserving privacy. Unlike traditional methods that require extensive prior knowledge analysis, DP-ARST simplifies the data preprocessing by only requiring the number of trajectories and the area's size. Furthermore, to better preserve the realistic characteristics of trajectories, we introduce motion constraints to guide the synthesis process. Comparative experiments demonstrate that DP-ARST achieves superior utility while maintaining robust privacy protection compared to existing synthetic trajectory methods.