Vehicle Trajectory Data Publishing Mechanism Based on Differential Privacy
Shanbin Li, Zhiyin Qi, Qinxue Li · 2021 China Automation Congress (CAC) · 2021
In recent years, the intelligent connected vehicle technology has been developing rapidly, which will vigorously promote the construction of smart city. Vehicle trajectory data is of great value for applications such as city services and data mining. However, it will greatly endanger the safety of participants if trajectory data is released unprotected. In this paper, we consider the velocity factor and propose Douglas Peucker with velocity trajectory compression algorithm (DPWV) to achieve trajectory compression with lower error. Then, a ω-trajectory perturbation mechanism based on differential privacy (ω-TPDP) is proposed to protect the sensitive information of trajectory data. Meanwhile, in order to balance the utility and privacy of trajectory data, a privacy budget allocation method is adopted based on the weight value and the risk coefficient of privacy leakage of the current location. Through theoretical analysis and experimental comparison, ω-TPDP can obtain commendable data availability under the premise of ensuring the privacy of trajectory.