Privacy and Utility Preserving Trajectory Data Publishing for Intelligent Transportation Systems

Xiangwen Liu,Yuquan Zhu

doi:10.1109/access.2020.3027299

Abstract

Nowadays, the extensive collection and storage of massive personal GPS data in intelligent transportation systems every day provide great convenience for trajectory data analysis and mining research, thus bringing valuable information for real-life applications. Yet, protecting personal privacy is also more challenging in the smart environment. When trajectories of individuals are published together with their sensitive attributes such as disease, income etc., one can use partial trajectory knowledge for identity, sensitive locations, and sensitive values of target individuals. We present $(\alpha, K)_{L}$ -privacy model and an anonymization scheme aimed at ${I}$ dentifying and ${E}$ liminating ${V}$ iolating privacy ${S}$ ubtrajectories (IEVS), to prevent privacy disclosure while preserving the accuracy and high quality of published trajectories. In particular, IEVS employs three anonymization techniques, i.e., trajectory splitting, location suppression, and sensitive value generalization to eliminate all subtrajectories violating $(\alpha, K)_{L}$ -privacy principle. Experiments show our scheme is effective to improve the data utility of anonymized trajectories when compared with previous work.

Highlights

The rapid development of the Internet of things (IoT) and big data technology has spawned many new smart application domains for the urban environment
To address the above problems, this paper aims to design a novel trajectory anonymization scheme in a combined data publishing scenario, where trajectory data without modification are released with sensitive attributes
We address the problem of privacy preservation in a combined data publishing scenario, where the trajectory data of an individual is published together with his/her sensitive attribute

Summary

Introduction

The rapid development of the Internet of things (IoT) and big data technology has spawned many new smart application domains for the urban environment. The intelligent transportation system (ITS) [2], for example, one of the important application domains in smart cities, generates great amounts of realtime GPS data every day Such abundant spatiotemporal information, organized as trajectories, reflects the historical traffic conditions of a city. If a published trajectory dataset includes all the trajectory information of this day, and there exists a trajectory containing the location-time data of the target’s visit to the gourmet shops, the adversary can associate this trajectory with the target accurately. He can further obtain the target’s sensitive information contained in the trajectory, such as the home address, travel habit, health condition, personal interest, etc.

Objectives

Methods

Conclusion