Achieving Privacy Preservation In Location Based Services

With the current evolvement of mobile devices and wireless communication technolo-gies,location-based services(LBSs)have been enjoying growing popularity in recent years.On the downside,the location disclosure when LBS are provided often implies sensitive per-sonal information such as one's health status,political beliefs and social relations,thereby raising severe privacy concerns.Therefore,how to preserve location privacy has become a key issue in LBSs.This paper aims to protect users' location privacy against spatio-temporal correlation attacks,location injection attacks,inference attacks,prevent privacy disclosure in indoor localization,and quantify performance of privacy preserving techniques in continuous LBSs and single LBS.Specifically,the innovations in this paper are as follows:1.RobLoP is the first work aiming at preserving location privacy against spatio-temporal correlation attacks in a uniform way,MMB and MAB attacks together are referred to as spatio-temporal correlation attacks.Most existing work only focuses on the MMB attacks.One attempt against MAB attacks leads to gratuitous "false-negative" cases and therefore cannot completely protect users' privacy.To this end,RobLoP converts such spatio-temporal correlations to geometric constraints in a uniform way such that the candi-date users cloaked together with user of interest can be identified practically.As a result,attackers cannot distinguish the location of user of interest.2.This paper presents ILLIA that enables k-anonymity-based privacy preser-vation against location injection attacks(LIAs)in continuous LBS queries.Existing studies characterize LI As and demonstrate LI As ' effectiveness(on disclosing users' loca-tion privacy).Unfortunately,they do not provide a corresponding solution to protect users'location privacy against LIAs.ILLIA is a general solution against LIA,without requiring in advance knowledge of how fake locations are manipulated.Moreover,it is scalable,s-ince it is robust to the number of users and light-weight in terms of the processing time.Furthermore,ILLIA is able to maintain a better trade-off between QoS and location privacy preservation.3.This paper proposes W3-tess synthesizes privacy preserving traces via enhanc-ing the plausibility of synthetic traces with social networks.Heuristic algorithms are first proposed and cannot model users' mobility behavior.The follow-up studies only consid-ered temporal and spatial behavior of users' mobility,completely ignoring the social be-havior of users ' mobility.Therefore,these work is susceptible to social relationship based location inference attacks.In contrast,both location privacy preservation and trace data u-tility guarantee(a.k.a QoS)are theoretically guaranteed in W3-tess.Moreover,W3-tess is applicable to most geo-data analysis tasks as long as they are composable.Furthermore,users are informed of QoS whenever they set up W3-tess.4.P3-LOC is the work with provably guaranteeing both users' location priva-cy and the localization server's data privacy.Existing literatures are algorithm-driven,cannot completely guarantee the localization server(LS)'s data privacy,and are encryption-based.P3-LOC is largely distributed and scalable,and can be applied to any localization system that complies with the two-stage localization paradigm.5.This paper proposes the first theoretical foundation that quantifies the perfor-mance of k-anonymity based techniques.Previous studies only rely on a number of exper-imental examples to validate k-anonymity's performance,and thereby lack of a theoretical guarantee.The theoretical foundation takes into consideration of adversaries' background information and gives non-asymptotic bound on the performance of k-anonymity based tech-niques.In addition,it bridges the gap between design and evaluation,enabling a designer to construct a more practical k-anonymity technique in real-life scenarios.