Research On Privacy Preserving Technique In Trajectory Data Publishing

In recent years, with the growing popularity of location technologies and location-based applications, application providers and research institutions accumulate a large number of trajectory data for research, analysis, and publishing. Trajectory data publishing is releasing trajectory data to academic and commercial research institutions used for urban planning, behavior analysis, business decisions, and other research, with the purpose of exploring trajectory data’s academic value and commercial value. Privacy preserving issues of trajectory data publishing have a critical significance in its development, and have become a hot research topic. Due to the characteristics of trajectory data such as large-scale, high-dimension, and rich-background, the research on privacy preserving issues is facing severe challenges.This thesis makes a deep research on the privacy preserving technology in trajectory data publishing, propose a (k, Δ)-anonymity model for trajectory anonymous problem, and based on this model we design and implement a trajectory privacy preserving algorithm named UPG. In the (k, Δ)-anonymity model, the inherent uncertainty of trajectory data is exploited, to reduce translation cost in the anonymous process. In the UPG algorithm, we first partition trajectories into line segments based on the MDL principle, and then anonymize these segments with cluster-constraint strategy. UPG algorithm can solve the problem of lack of diversity in the anonymous group of traditional trajectory privacy preserving algorithm, and effectively prevent re-clustering attacks against the characteristics of publishing data. Finally, in the section of experiment, we assess the performance of our privacy preserving algorithms, respectively in terms of the data quality and data efficiency, and compare our (k, Δ) algorithm with classic NWA algorithm. As result shows, privacy preserving algorithm based on the (k, Δ)-anonymity model has a smaller anonymous cost in most case. And UPG algorithm can greatly improve the level of privacy preserving, with a minimum data quality cost.This thesis first introduces the relevant background knowledge of privacy preserving technique in trajectory data publishing. Then, we research on the issues of trajectory anonymous model in trajectory privacy preserving, and propose a trajectory anonymous model based on the inherent uncertainty of trajectory named (k, Δ)-anonymity model. Then, with the (k, Δ)-anonymity model, we propose a partition group based algorithm for trajectory privacy preserving named UPG. Finally, we summarize this thesis, and point out the existing problems and future research direction of privacy preserving.