Research On Location Privacy Preserving Technology Based On K-Anonymity In Location Service

With the rapid development of mobile Internet technology and the popularity of smart phones,more and more people start to use Location-Based Service(LBS).People could provide their locations and get corresponding services by LBS.For example,users can find the friends nearby,etc.While LBS provides convenience for our daily life,the issue of personal privacy leakage should not be neglected.Although scholars have proposed a variety of effective location privacy protection strategies,most of them focus on the stage of providing location to LBS server.There is few privacy protection policy that focus on the stage of getting location from Location Provider(LP).Getting the user's current position is the premise of using LBS.However,when users use the network location service provided by LP,they usually need to provide the current location fingerprint information,which could also lead to the privacy disclosure of the user's location.Therefore,this paper focuses on the location privacy protection in location service.Firstly,we design the OFC(Optimal Fingerprint Construction)algorithm to realize k-anonymity to protect location privacy in location service by adding k-1 dummy location fingerprints in the original location request.The key to this strategy is to generate dummy location fingerprints that LP can not distinguish from real location fingerprints.In order to solve this problem,an anonymous knowledge base based on weighted graph model is designed,which could indicate the spatial distribution of access points(AP).OFC is used to construct a dummy location fingerprint by selecting the nearest APs to ensure it has similar performance with real location fingerprints in positioning process.Secondly,considering that the attacker may use the auxiliary information to judge the rationality of the location estimated from dummy location fingerprint,we improves the OFC algorithm and designs the SOFC(Score-based Optimal Fingerprints Construction)algorithm and COFC(Continuously Optimal Fingerprints Construction)algorithm.The SOFC algorithm considers the access frequency information that the attacker may use.The algorithm gives a reasonable score for each AP according to the historical access,and select the AP whose score is near the AP in the real fingerprint while generating dummies.The COFC algorithm takes into account the trajectory characteristics of attackers who may make use of the continuous location requests of users.It ensures that the trajectories of dummies and the real trajectory of users have higher directional similarity and distance similarity.Finally,we compared our proposed algorithm with the existing location privacy protection algorithm.The experimental results show that the OFC algorithm has a high guarantee for anonymous quality and significantly reduces the time consumption.SOFC algorithm and COFC algorithm have better privacy protection effect for attackers with auxiliary information while guaranteeing anonymous quality,but the time consumption is more than that of the OFC algorithm.