Research On User Context Information Privacy Protection Framework And Mechanisms For Location-Based Services

With the growth of a rich variety of information technology and extensive use of mobile social networking platforms,location-based service(LBS)has recently become popular in our everyday lives.In addition,the widely use of some emerging technologies such as social networks,mobile Internet,and big data have further increased the usage and interaction between mobile users and LBSs,as well as improved the experience of users in LBS,but it would comes with a very high cost to user privacy.When users conveniently access various LBS applications,they will inevitably leave a large number of digital traces and service attributes in the network.However,the user context attached to these service attributes or digital traces can reveal other sensitive personal information,such as their physical condition,personal interests,habits,and relationships.Therefore,the disclosure of the personal information to untrusted third-party adversarial entities often results in user's privacy breaches.A large number of location privacy-preserving mechanisms(LPPMs)have been presented in the literature to try to deal with the location privacy issues.However,a majority of existing approaches relies on different technical means in the prevention of the disclosure of user privacy information,and thus has advantages and limitations.In order to address these shortcomings and limitations,we devise a privacy-preserving mechanism based on game theory to protect users against private information leakage arising from malicious third-party observers.Essentially,it is a pivotal for the privacy protection to take a balance between the amount of information a user is comfortable to share with others and the amount of information an adversary can learn in user's private information.In particular,we first present the current solutions on privacy protection mechanism,and analyze their features and limitations.In contrast to existing solutions,we investigate new approaches to study privacy protection problems from the perspective of behavioral interest and game-theory,and then establish a privacy protection mechanism in the process of visitors accessing the users'private information.We also model a scenario of privacy protection strategy as a repeated game context that derives the situation of the prisoner's dilemma game between third-party visitor and the owner of the private information.In this process,the strategies adopted by both sides and the corresponding profits are analyzed to establish a game model,which is the basis for realizing the protection of user's private information,and finally achieves the purpose of effectively protecting the location privacy the user.We evaluate the corresponding incomes to both sides using a point of view that allows the visitor to access a certain amount of user's private information and denies further access when disclosure of privacy is forthcoming.We also conduct some experiments to demonstrate the superiority and effectiveness of the proposed game-theoretic model as well as its advantages over other existing solutions through comparison analysis.Furthermore,we present a privacy protection framework to study the interactions between location privacy,service quality loss,and resource utilizations in LBS systems.The developed protection model combines the various pieces of information about users,their access pattern to LBSs,and their quality of service and location privacy requirements.The protection mechanism is evaluated along with other popular solutions such as,precision-based and dummy-based LPPMs using a real-world datasets.The advantage of this study is to be able to decrease the precision of the exposed locations or generate dummies locations.The systems are evaluated considering the level of privacy protection,the quality of service provided to the LBS user and energy consumption analysis.The experiments indicate that dummy-based LPPMs provide better achievable privacy under a given combination of energy and quality-loss constraints,and once a certain amount of privacy is reached both the dummy-based and precision-based LPPMs only perturb the exposed locations.Theoretical analysis and experimental evaluation both show that this thesis has a step towards contribution for modeling,design and evaluation of the location privacy enhancing mechanisms.