Research On Privacy Protection Technology Of Differentially Private Trajectory Sharing

In recent years,with the development of mobile devices and communication network technology,location based services(LBS)have covered all aspects of national economy and social life.At the same time,the widespread use of location services will generate a large amount of trajectory data,which contains rich spatiotemporal information.Reasonable analysis and publication of these trajectory data are of great value for many applications.However,unauthorized direct publication or sharing of users’ personal trajectory information may lead to serious privacy breaches.Therefore,how to use trajectory data while protecting user privacy has become a key challenge.The user trajectory privacy protection in location-based services is mainly divided into two scenarios: publishing trajectory privacy protection and real-time trajectory privacy protection.This article proposes corresponding privacy protection schemes around the privacy leakage and insufficient trajectory utility issues in the two trajectory privacy protection scenarios.The main research of this article is as follows:(1)Most of the existing privacy protection methods for publishing trajectories are based on differential privacy technology to generate synthetic trajectories that are highly similar to real trajectories.However,this type of method can result in low availability of synthetic trajectories and susceptibility to location privacy inference attacks.This chapter proposes a utility enhanced synthesis of differentially private trajectories(Uti E-DPT)method.This method divides the real trajectory into multiple spaces,builds a fine-grained adaptive density grid to discretization the real trajectory,and designs a differential private Markov mobility model,a trajectory travel distribution model,and a trajectory length distribution model suitable for the adaptive density grid to more accurately extract the key statistical characteristics of the utility maintenance of the real trajectory.At the same time,in order to prevent trajectory privacy leakage,differential private perturbation is applied to the extracted features.Finally,based on the extracted features and the anti attack constraint strategy,a synthetic trajectory is generated to resist inference attacks.The experimental results on simulated and real datasets show that compared with existing trajectory synthesis privacy protection methods DP Star and Ada Trace,Uti E-DPT enhances the availability of synthesized trajectories while protecting trajectory privacy and resisting location privacy inference attacks.Without resisting inference attacks,the query error of Uti E-DPT in generating synthetic trajectories is reduced by 21%-26% compared to Ada Trace and 31%-51% compared to DP Star;After resisting inference attacks,although the robustness of Uti E-DPT in generating synthetic trajectories is reduced by about 1% compared to Ada Trace,the query error is reduced by 15% to 20% compared to Ada Trace,achieving a better balance between privacy protection and utility.(2)The existing real-time trajectory privacy protection methods have problems such as insufficient trajectory privacy protection,time-consuming and laborious construction of query service databases,and low query accuracy in requesting services.This chapter proposes a real time trajectory privacy(RTTP)method based on differential privacy and homomorphic encryption.This method divides regions based on historical trajectories,and then crawls the interest points within the region to construct a region interest point cache.When users send service requests,Laplace noise is added to location points using differential privacy combined with geographic indistinguishability.At the same time,the Paillier algorithm is used to encrypt the query content,and the server side utilizes the homomorphic characteristics of the Paillier algorithm to respond to service requests in a ciphertext state.Protect trajectory privacy and query content privacy without affecting service quality.Adding fixed noise to location points can cause significant errors and lead to a decrease in service quality.RTTP personalized adds Laplace noise by calculating the privacy level of location points.At the same time,in response to the problem of duplicate requests for services at adjacent location points in the trajectory,if the distance between the two requests for services is similar,the perturbed location of the previous location is used to request services,reducing the time and computational costs of the current service request.