Research On Differential Privacy Preserving In Trajectory Data Publishing

In recent years,with the popularity of positioning technology and location-based applications,a large number of user trajectory data have been accumulated.In order to explore academic and commercial value of trajectory data,the accumulated user trajectory data have been published to academic and commercial organizations for research on urban planning,behavioral pattern analysis,business decision making and so on.The privacy protection within the releasing of trajectory data has become a research hotspot.Because of trajectory data features in high dimension,temporal and spatial correlation,and rich background knowledge,the privacy of trajectory data are faced with different challenges from traditional privacy protection.For the trajectory data characteristics,differential privacy protection method is applied.Firstly,a uniform sampling sphere model is proposed,based on the model,the global noise algorithm and the partial noise algorithm are implemented,and it is proved that both of them conform to the generalized(?,?)differential privacy protection.In order to further realize strict differential privacy protection,in aspect of vector,Previous Position Vector algorithm,PPV,is presented,in which random noise in dimensions of both distance and angle are added to achieve differential privacy protection.In the final experimental section,the performance of PPV algorithm is evaluated from the metrics of error and stability.After process of differential privacy protection,the published trajectory data can protect the user's personal privacy and ensure the availability of trajectory data as well.The results show that PPV algorithm has high degree of privacy protection and good performance.