Privacy Protection Methods For Location Based Service

The increasingly popular Location Based Service(LBS)brings great convenience to people's lives.However,while enjoying high-quality services,users also face the risks of privacy leakage.Therefore,how to avoid location privacy leakage has become a key issue restricting the development of LBS.It is of great significance to research the characteristics of location privacy disclosure and protect users location privacy.This thesis focuses on the single location privacy protection,continuous location privacy protection in real-time LBS and trajectory location protection in offline LBS,the main work includes the following three aspects:(1)In order to solve the problem of single location privacy leakage in real-time LBS.this thesis has proposed the real-time single location privacy protection scheme based on anonymous entropy.In particular,this thesis proposes an anonymous entropy method for constructing anonymous regions.The regions to be protected are divided into densely populated regions and sparsely populated regions,and an anonymous region construction algorithm based on K-anonymous(K-DDCA)in densely populated region and an anonymous region construction algorithm based on K-anonymity(K-SDCA)in sparsely populated region are proposed respectively for location privacy protection in their respective regions.On the one hand,the K-DDCA algorithm constructs anonymous regions based on the distance between requesting users and the difference of requesting content.On the other hand,in sparsely populated region,since the k-1 neighbors cannot be found,this thesis proposed K-SDCA algorithm generating k-1 dummy users based on historical records and geospatial distribution to construct anonymous region together with real user.(2)In order to solve the problem of continuous location privacy leakage in real-time LBS,this thesis has proposed the real-time trajectory privacy protection scheme based on prediction mechanism.The user trajectories are represented in the form of grid,and then a trajectory similarity metric method using user's behavior pattern and location reachability constraint is proposed.The proposed position prediction based on similar trajectories algorithm(PPST)searches for similar trajectories to predict the location of the target user at the next moment,and the predicted location is assigned a risk level.According to the different risk levels,The enhanced dynamic anonymity protection algorithm(EDAA)is further proposed to dynamically protect the location of different risk levels.(3)In solving the problem of location privacy leakage in offline LBS,that is,location privacy leakage in the trajectory publishing.This thesis improved the dummy trajectory rotation generation method and then proposed the offline trajectory privacy protection scheme based on space-time constraints.Firstly,the selection of the pivot point and the overall motion direction of the trajectory are constrained to prevent the sensitive position from being selected as the pivot point and ensuring that the overall motion direction of the trajectory is similar.Secondly,The initial dummy location generation algorithms based on spatio-temporal constraints algorithm(IDG)and a trajectory generation algorithm based on space-time constraints algorithm(TGC)are proposed to generate the adjacent locations under the pivot point and the other locations under the non-pivot point.These ensure that the generated dummy trajectory has the same motion pattern as the real trajectory.Finally,a large number of simulation experiments carried out based on the proposed schemes.The experimental results show that these schemes can achieve better privacy protection and higher efficiency than the existing schemes.