Research On Differentially Private Location Privacy-preserving Technology In Location-based Services

Currently,China has entered the era of Internet of things.With the rapid development of mobile network and communication technology,the popularity of location-based services(LBS)has greatly enriched all kinds of travel and social ways of users.Nevertheless,because users need to send their current location information and location-related query content to untrusted location-based service providers(LSP)when using various LBS applications,this may cause serious privacy concerns.The user's concern about personal location privacy security has become one of the important factors hindering the healthy development of LBS industry.Therefore,the research can not only protect the location privacy of users,but also provide high-quality service for users.The method of location privacy protection is of great significance to the healthy development of LBS industry.Differential privacy has become the most popular protection technology for current location privacy protection because of its advantages of security,practicability and verifiability.However,the existing location privacy protection methods based on differential privacy often focus on the improvement of privacy protection,ignoring the better trade-off between privacy protection intensity,service availability and computing overhead.In addition,this method is mainly aimed at the privacy protection of a single location point,which can not meet the needs of the privacy protection of the location trajectory,and the process resource cost of generating the simulated trace is large.Aiming at the above problems,this paper introduces Semantic location,Density-aware grid,Markov chain and other means,combined with differential privacy mechanism to carry out research,and provides a location privacy protection mechanism that can solve the above problems.The main research work of this thesis is as follows:(1)This thesis analyzes and compares the existing location-based differential privacy protection methods,points out the problems of the existing methods.Through in-depth investigation and research,this thesis studies and designs specific and effective solutions.(2)Aiming at the problem of more reasonable noise addition in location privacy protection,a differential private location privacy-preserving scheme with semantic location is examined in this paper,which can solve the contradiction among privacy-preserving,service availability and time overhead.The proposed method firstly constructs the expected distance by employing the framework of geo-indistinguishability,then determines the sensitivity of different locations by using the privacy quality function and requirement function,and finally adds Laplace noise to different types of region at fine granularity according to the location sensitivity.The experimental results and analysis show that the proposed method can reduce the time cost by up to 9.6% compared with the existing method.Compared with the classical Planar Laplace(PL)mechanism,the maximum protection intensity in Point of Interest(Po I)sparse area can be increased by up to 50%;Compared with Elastic Mechanism(EM),the quality of service for Po I dense area can be increased by 5.3% on average without increasing the time cost,and its stability when providing protection for sparse areas without losing the existing privacy protection strength can increase by 41.78% on average.Thus,the proposed method can effectively protect location privacy without increasing computing overhead,and obtain a better balance between service quality and privacy protection.(3)In order to solve the problem that the existing location trace protection mechanisms based on differential privacy seldom consider the availability of generated trace and the security of location privacy.A simulation trace generation mechanism based on AdaTrace framework combined with Density-aware grid and Markov chain is proposed.The scheme preserves the relevant key practical information in the real trace,and combines the Laplace noise addition mechanism to make the generated simulated trace not only able to resist the inference attack with background knowledge,but also has high availability and low computation time cost.The results of 24 experiments with 8 indicators in three directions show that the proposed method can achieve the tradeoff among privacy protection,trace utility and time cost on the premise of ensuring the availability of generated trace,so as to achieve more personalized location trace privacy protection.