Research And Application Of Privacy-preserving Method For Uncertain Trajectory Data Of Moving Object

With the development of global navigation and positioning system,plenty of trajectory data appear.However,trajectroy data include much personal information so that mining trajectory data maybe lead to privacy disclosure.There are many related works which assume the data are precise.However,device collecting data and data delayed updating result the uncertainty of trajectory data.So the methods based on certain data can not process the uncertain trajectory data well.Therefore,to the uncertainty of trajectory,we present methods which apply to uncertain trajectory data.Nowadays,there are two types privacy-preserving method on data publishing,the one is based on K-anonymity,the other is based on Differential Privacy.This thesis present two methods to privacy-preserving of uncertain trajectory data publishing based on above-mentionedies,Restore Its True(RIT)and Reconstruct Trajectory based on Difference of privacy(RTD)privacy-preserving methods.RIT method based on K-anonymity.It solves two drawbacks of(k,?)-anonymity model that RIT method reckons the uncertain radius ? as a variable;it translate the linear model to area model.First,it use the method in probability statistics to model the trajectory to an uncertain area.Second,it put the similar uncertain area into acluster and sanitize them in an equivalence class.Finally,it use(k,m)-prefix treealgoRIThm to anonymize.RTD method based on Differential Privacy.Firstly,It adds Laplace noise to sampling points within uncertain radius and let them satisfy RTD-protective degree.And then it clusters the trajectories method which have been noised with dynamic programming.To reduce the information lossing,it does randomized reconstruction.Finally,we design and implement a prototype system to integrat an apply the two methods.We use the system to experimentalize in the real dataset.The experimental results express that RIT method can restore the authenticity of data effectively so that the usability is improved and it can preserve the provacy;RTD method can reduce the information lossing of data observably.In the way of operational efficiency,it is better than RIT method.