Font Size: a A A

Research On Location And Trajectory Privacy Protection In Mobile Sensing Computing

Posted on:2015-12-19Degree:DoctorType:Dissertation
Country:ChinaCandidate:S GaoFull Text:PDF
GTID:1108330464968880Subject:Computer system architecture
Abstract/Summary:PDF Full Text Request
With the popularity of the Mobile Internet, mobile sensing applications such as Participatory Sensing and Location-based Services have attracted extensive attention. In these applications, a user location and trajectory regarded as quasi-identifiers deeply reflect his/her sensitive information, such as home address, health condition, life style, and so on. To make the user be willing to share his/her collected local knowledge and use various application services related to location, the research on the user location and trajectory privacy has important implications for ensuring his/her personal safety and promoting the health development of Mobile Internet.This dissertation studies the location and trajectory privacy protection in the process of data collection and data use, which includes a mechanism to achieve the tradeoff between location privacy and quality of services and trajectory privacy protection in participatory sensing; a framework to protect trajectory privacy for participatory sensing; a personalized anonymization model to balance trajectory privacy and data utility; a method to protect location privacy for group nearest neighbor queries in location-based services. The main research achievements are as following:1. A mechanism to achieve the tradeoff between location privacy and quality of services and trajectory privacy protection in participatory sensing is studied. For those participatory sensing applications which require precise user locations, a more practical approach on the basis of dummy location technique to protect location privacy and meet the requirement of quality of services is proposed. On this basis, an approach which uses to construct several trajectories that are similar to that of the user is used to protect the user trajectory privacy. Then, an evaluation framework is introduced and a method to measure the similarity relationship between the user’s trajectory and the constructed trajectories is given. Finally, simulation results on the basis of the evaluation framework prove the effectiveness of our approach.2. Based on the analysis of current trajectory privacy protection techniques, a trajectory privacy-preserving framework TrPF for participatory sensing is proposed to protect the linkage relationship between the user’s identity and his/her trajectory from being leaked. In TrPF, an improved theoretical mix-zone, pseudonym technique and-anonymity technique are combined to localize trajectory privacy protection and achieve the unlinkage between the user’s identity and his/her trajectory by only dealing with the sensitive trajectory segments. Compared with the existing trajectory protection methods, the proposed scheme can achieve the trajectory privacy protection while effectively reducing the information loss and storage cost.3. Considering the tradeoff between user trajectory privacy and data utility, according to the user’s dynamic requirements in different application scenarios, a personalized trajectory anonymization model is proposed to select the trajectory-anonymity set that meets the demands of the user. Specifically, the anonymity process includes trajectory pre-processing, trajectory graph construction and trajectory-anonymity set selection. The relationship among trajectories can be abstracted as a corresponding graph model and the trajectory-anonymity set can be selected by the greedy algorithm from graph theory. Simulation results prove that our approach can provide a trajectory-anonymity set that meets the user’s personalized requirements.4. To achieve location privacy protection for group nearest neighbor queries in Location-based Services, under the distributed peer-to-peer system structure, according to the motion status of a user group, two approaches named location random perturbation and threshold secret sharing version of Paillier cryptosystem are proposed to compute the center location of the user group. The group nearest neighbor can be obtained on the basis of the calculated center location while the user group location privacy can be protected. Compared with existing related work, the proposed proposal can effectively resist against the existing distance interaction attack and collusion attack and achieve flexible GNN queries, while it costs lower network resources.
Keywords/Search Tags:Location Privacy, Quality of Services, Trajectory Privacy, Linkage, Relationship, Data Utility, a Personalized Anonymization Model
PDF Full Text Request
Related items