Research On LBS Protection Based On Game Theory And Differential Privac

Location-based service(LBS)provides great convenience for people’s life,but also has the problem of user privacy leakage.In response to the problem of privacy leakage in LBS,scholars at home and abroad have proposed a large number of location privacy protection schemes,but these schemes have the following problems:Firstly,the system architecture used in the scheme has problems such as single point of failure or performance bottleneck of mobile devices;Secondly,the scheme cannot effectively defend against single-point attacks and inference attacks and lacks the trade-off between service quality and privacy protection level;Thirdly,the scheme does not make reasonable assumptions about the attacker model,and its effectiveness depends on the attacker’s background knowledge,meanwhile the current location privacy protection mechanisms based on differential privacy all have certain defects.Aiming at the problems above,this paper designs related solutions based on k-anonymity,dummy,optimization problem,game theory and differential privacy,etc.The research work and innovation of this paper are as follows:(1)A k-anonymous location privacy protection scheme based on dummy and Stackelberg game.Aiming at the shortcomings of the system architecture of the current location privacy protection scheme and the inability to effectively resist single-point attacks and inference attacks,based on dummy,k-anonymity and Stackelberg game technology,a k-anonymous location privacy protection scheme based on dummy and Stackelberg game.The specific work of the scheme is as follows:Firstly,by analyzing the research status of the location privacy protection system architecture and based on the previous research work,an improved central server system architecture is proposed;Then,considering the characteristics of side information,location semantics,physical dispersion of locations,and combining k-anonymity,dummy technology,and offset location,etc.A dummy selection algorithm based on location semantics and physical distance has been proposed;Finally,the relationship between the user and the attacker is modeled as a Stackelberg game,and the Nash equilibrium of the game is solved by defining the relevant linear programming,so that the anonymity effect is further optimized.(2)A privacy protection scheme based on Stackelberg game and improved differential privacy.Aiming at the shortcomings of the current differential privacybased privacy protection mechanism,based on game theory,differential privacy,optimization problems and gradient descent algorithms,etc.we have designed a privacy protection scheme based on Stackelberg game and improved differential privacy.The specific work of the scheme is as follows:Firstly,we deeply analyze dissect the deficiencies of existing classical definitions of location privacy,on the basis of which,an improved definition of differential privacy(τ-DP)has been proposed;Finally,on the basis of τ-DP,we have analyzed the zero-sum Bayesian Stackelberg game between users and attackers in location privacy protection,constructs the optimal privacy protection mechanism from the trade-off between function and privacy,and introduces the gradient descent algorithm and cross entropy algorithm.An approximate optimal solution for solving nonconvex problems instead of the actual solution of the optimization mechanism.