Novel trajectory privacy protection method against prediction attacks

作者：

Highlights：

• Formally define the privacy attack method of trajectory prediction attack.

• A trajectory prediction mechanism is proposed to simulate prediction attack.

• A sliding window mechanism is designed to limit the privacy budget of trajectories.

• Publish perturbation trajectories that satisfy differential privacy.

摘要

•Formally define the privacy attack method of trajectory prediction attack.•A trajectory prediction mechanism is proposed to simulate prediction attack.•A sliding window mechanism is designed to limit the privacy budget of trajectories.•Publish perturbation trajectories that satisfy differential privacy.

论文关键词：Privacy protection,Trajectory data,Differential privacy,Geo-indistinguishability,Hidden Markov model

论文评审过程：Received 15 September 2021, Revised 25 July 2022, Accepted 17 September 2022, Available online 23 September 2022, Version of Record 29 September 2022.

论文官网地址：https://doi.org/10.1016/j.eswa.2022.118870