Research On Trajectory Data Protection Method Based On Differential Privacy

With the rapid development of mobile intelligent devices,more and more trajectory data are generated,which are collected by third-party service providers and then stored in the database.Trajectory data mining and analysis can solve many problems.However,trajectory data usually contains a lot of personal sensitive privacy information of mobile objects,which can be used directly without any protection,Personal sensitive privacy information of mobile objects will be leaked,which will cause huge losses to the mobile objects themselves,and even bring personal threats.Therefore,the privacy of trajectory data has attracted the attention of many researchers at home and abroad.In the existing research,there are mainly two methods to protect trajectory data,by using differential privacy:the first is "Local Noise",the second is"Global Noise".This paper analyzes the advantages and disadvantages of the two methods,and proposes a new method of trajectory differential Privacy Preserving Based on local noise;Furthermore,a differential privacy protection method based on location distance is proposed for a real-time interactive service request application scenario between mobile objects and servers.The main research contents are as follows:1)To solve the problem of low data availability of the existing trajectory data protection methods based on differential privacy,this paper proposes a new trajectory differential privacy protection method based on fuzzy c-means clustering(Fuzzy C-means Based Trajectory Differential Privacy Preserving,FCMBTDPP).Firstly,the stay points in the moving object trajectories are selected,that is,the positions with longer staying time.Then,the remaining positions of trajectories are clustered by fuzzy c-means.Finally,Laplace mechanism is used to add noise to the stay points and clustering centers.This method only adds noise to some position points of trajectory data,which can protect the privacy of trajectory data and improve the availability of data.2)In the process of real-time interaction between mobile objects and servers,mobile objects can obtain the desired services by uploading their own location data.In the process of interaction between mobile objects and servers,three application scenarios are static,low-speed and high-speed.Aiming at the problem that the privacy budget is accumulated with the increase of location points in trajectory data,a differential privacy protection method based on location distance is proposed,(Trajectory Differential Privacy Preserving based on Location Distance,TDPPLD).First,by judging the distance between the current location and the previous location,select the location point to be uploaded to the server,then add noise to the location point and then upload it to the server for service request.Finally,the experiment verification is conducted in the Geolife data set(low-speed operation scenario)and T-drive data set(high-speed operation scenario),it is proved that this method can not only protect the privacy information of track data,but also improve the availability of data.