| With the development of mobile computing,the Mobile Crowdsensing(MCS)technology based on wireless sensor network and crowdsourcing is emerging.However,there are many location privacy concerns in the life cycle of a MCS activity.For example,in the task assignment phase,the location data of the participants are required to optimize the task assignment algorithms.In the sensing data collection phase,strong spatiotemporal correlations among the sensing data lead to exposure of participants’ trajectories.To solve the above problems,two phases location privacy protection methods were proposed in this thesis.Moreover,based on the above researches,a student attendance management method with privacy protection feature is proposed to protect students’ privacy.The main research contents of this thesis include the following three aspects:(1)A Location Privacy Preservation Method in the Task Assignment PhaseAiming at the location privacy protection problem of participants in the task allocation phase of MCS activities,this thesis proposed a Task Allocation Method based on Geo differential privacy and distortion privacy(TAMG).First,in order to ensure the indistinguishability of participants’ location and the controllability of attackers’ inference errors,TAMG needs to generate the location obfuscation matrix on the MCS platform according to the differential privacy and distortion privacy constraints.Second,in order to implement TAMG in a non-trusted server framework,TAMG deploys the location obfuscation operation on the participants’ mobile devices,and participants upload their obfuscated locations to the MCS platform.Finally,the MCS platform optimizes task allocation according to the obfuscated locations.Moreover,this thesis proposes an optimization method TAMG&GA based on genetic algorithm for the local optimal solution problem of TAMG.This thesis conducts experiments on the real-world dataset,and counts the time consumption of TAMG and the expected average travel distance ATD of the participants.The results show that TAMG can limit the expected average travel distance to a small value while satisfying high level privacy constraints.Compared to TAMG,the optimization method,TAMG&GA can reduce the ATD value by up to 26.4%.(2)A Trajectory Obfuscation Method in the Sensing Data Collection PhaseAiming at the participants’ location privacy protection problem in the sensing data collection phase of MCS activities,this thesis proposed a Location Privacy-preserving Method based on Trajectory obfuscation(LPMT).First,LPMT extracts the stay points of trajectories as the characteristics of the trajectories by using the slide window algorithm.Second,the exponential mechanism is used to generate the target obfuscation subregion for each stay point,and the utility calculation method combines the Euclidean distance and location context similarity is implemented to balance the obfuscation quality and the data quality loss.Finally,the Laplace mechanism is implemented to sample GPS points in the target obfuscation subregion and upload the sampled GPS points to the MCS platform.This thesis conducts experiments on the real-world dataset,and calculates the obfuscation quality and data quality loss of LPMT.The results show that,compared with the baselines,LPMT can reduce the data quality loss by more than 20%while providing the same level of obfuscation quality.(3)A Privacy-Protected Classroom Attendance Management MethodAiming at the privacy protection problem in the student attendance management systems,this thesis proposed an Attendance Management Method based on Crowdsensing(AMMoC).AMMoC includes two phases.In the initialization phase,students submit their logical location in the classroom according to the classroom seating map.The verification phase includes two modules,task assignment module and location verification module.In the task assignment module,AMMoC finds the optimal subregion sequence and verifiers through the Monte Carlo Tree Search(MCTS)algorithm,then sends verification tasks to the verifiers.In the location verification module,AMMoC verifies the authenticity of the student’s logical position according to the results of the verification tasks.The experimental results show that the AMMoC system has the advantages of well privacy protection,short attendance time and high accuracy. |
