Trajectory Privacy Protection Method Based on Differential Privacy in Crowdsensing
Qiong Zhang, Taochun Wang, Yuan Tao, Fulong Chen, Dong Xie, Chuanxin Zhao
Abstract
With the widespread popularity of smartphones, watches, and other devices, mobile crowd sensing has garnered significant public attention. Application service providers publish crowd sensing tasks, and users actively participate in collecting relevant sensing data, which are then submitted to servers. However, these data contain users’ personal privacy. Therefore, this article proposes a trajectory privacy protection method based on differential privacy(CTDP). First, the article conducts clustering based on the features of user trajectory data to extract feature regions of the user trajectory. Then, a personalized privacy budget allocation method is developed based on the number of trajectory points in the feature region and the user's privacy requirements for sensitive trajectory points. A set of confusion points is generated within the feature range and a score is calculated based on its similarity to the trajectory points. Subsequently, the sampling probability is calculated based on the score and privacy budget of each confusion point, and finally the confusion points are selected through random sampling. The internationally recognized real dataset Cabspotting data was used for experimental evaluation. The experimental results indicate that the method proposed in this article exhibits excellent performance in terms of data availability while providing sufficient privacy guarantees.