The geograhic locations of users offered by mobile devices are of great value for crowdsensing applications,such city planning,intelligent transportation system,etc.However,it will leakage users' privacy.The traditional privacy protection models(e.g.,k-anonymity)can not provide sufficient privacy protection because of its vulnerability to background attack.Differential privacy,as a formal framework in statistical database,can quantify individual privacy and resist background attack.We study the protection methods in location privacy based on differential privacy.The main works include:(1)In order to protect the location privacy and encourage users to check in,an incentive mechanism with personalized privacy protection is proposed.First,we perturb the time and location information in the check-in data by using the differential privacy method.We then design a reward strategy based on data quality to increase the quality of check-in data,which calculates the reward amount according to the check-in data quality of check-in users.The simulation experiments indicate that the proposed method can preserve users' privacy effectively,at the same time better guarantee the quality of check-in data.(2)A differentially private spatial decomposition based on Staircase mechanism is proposed.We first investigate the relationship between non-uniform error and query intersection area,and utilize the linear least square to fit the linear relation between them.Then we deduce the optimal partition granularity by minimizing non-uniform error and noise error.In the experiments,we use two real world datasets to evaluate the performance of the proposed method.Experiments show that the proposed two-dimensional spatial publishing method makes a good trade-off between data privacy and utility.(3)In order to improve the performance of trajectory releasing,a trajectory merging publication method based on Staircase mechanism and k-means|| clustering is proposed.wefirst propose two trajectory merging schemes based on k-means|| clustering.Afterwards,we propose a bounded Staircase noise generation algorithm.Theoretical analysis and experimental comparison show that our proposed publication methods significantly outperform existing approaches in terms of data utility and efficiency,while preserving differential privacy. |