Research On Privacy-preserving Method For Location Data Query

The arrival of the era of big data and the vigorous development of the Internet have spawned a large number of moving object trajectory data.Research and application based on trajectory data play an important role in urban planning,user behavior analysis,and frequent pattern mining etc,which have a great impact on people's production and lifestyle.However,the trajectory data contains abundant information of mobile users in the spacetime dimension.Releasing personal trajectory data and location statistics directly will reveal the private information of users.The existing trajectory publishing algorithms mainly rely on batch processing platforms,and pay little attention to real-time privacy protection processing in streaming scene.It is particularly difficult to achieve real-time privacy protection processing on the trajectory data stream due to its own characteristics such as high speed,mass,and uncertainty.Moreover,in the current statistical histograms publishing based on location data,real-time privacy protection support is rarely provided here.To solve the above problems and challenges,the privacy-preserving streaming trajectory publishing framework studies real-time query and privacy protection publishing based on trajectory data flow,which includes two parallel execution modules,namely streaming trajectory data publishing(Trajectory Streaming Publish,hereafter,referred to as TSP)and visitor count publishing(Visitor Count Release,hereafter,referred to as VCR).TSP is a novel module that publishes a streaming trajectory based on the privacy model !-.According to the user's personal trajectory query request,the synthesized trajectory segments with privacy protection are fed back in real-time,and ensure that they meet the personalized privacy preferences of different inquirers.Specifically,TSP divides the trajectory into segments and allocates privacy budget to each of them,and then returns new synthesized trajectories after tuple sampling and generalization processing.In addition,the VCR module can periodically publish statistical histograms about the distribution of location data.It includes two region visitor count histogram publishing algorithms !and !,and a histogram publishing algorithm " that supports grouping to better publish location statistics adaptively.Experimental results show that in the process of publishing streaming trajectories with privacy protection,a lower privacy budget will bring more noise to the output results,which will increase the deviation between the original trajectory and the synthesized one.Moreover,a dataset with a larger road network area is suitable for a larger time window size,and for a dataset with a smaller road network area and shorter trajectory distance,a smaller value of is the better selection.In addition,comparing with the traditional trajectory publishing algorithm N-grams,the module TSP achieves better publishing effects in terms of privacy protection and data availability.Comparing with the regional visitor counting histogram publishing algorithm !,the proposed adaptive grouping-based histogram publishing algorithm " has achieved a better accuracy of region visitor count effect under the same level of privacy protection.Finally,when choosing the appropriate parameter configuration,the proposed framework and its algorithms can effectively ensure privacy while achieving high data availability.