The Research On Privacy-preserving Matching Mechanism For Online Ride Hailing Service

In recent years,with the growth of mobile communication and sharing economics,the Online Ride Hailing(ORH)service has been developed explosively,and becomes a popular mean of travel.The service provider can apply real-time matching between riders and drivers over their locations,by which riders can improve the travel efficiency by making requests ahead of schedule,while drivers can earn the profit utilizing idle vehicle resources.However,ride-hailing service is a location-based service,the protection of user privacy is always ignored,the leakage of users’ location may expose users’ daily routines and identities,which may incur other serious consequences.With the arrival of the bigdata age,the amount of user data has been massively grown,the service provider can improve the user experience by utilizing user datas,but should protect the user privacy from the third party while providing service.The State Department pays high attention to regulation of data security,and promote its development as one of the country strategy,but even though,the problem of privacy leakage has been occurring frequently.Therefore,the power of regulation is limited,to protect user’s privacy,technique is needed to ensure that no other partiy can access users’ sensitive infomation;moreover,besides user privacy,system efficiency and user experience are also necessary,which should be considered at the same time.Based on the problems stated above,we design two privacy-preserving ride-hailing schemes under the scenarios of ride-matching and ride-sharing respectively.1.Privacy-preserving ride-matching scheme with prediction.Existing private ridematching schemes always match the nearest driver for the requesting rider,but it is not the optimal strategy from global perspective,which may cause the waste of the empty distance.To solve the problem,we propose a privacy-preserving ride-matching scheme with prediction.In the scheme,a rider request prediction model is introduced to predict the rider distribution in a short term.With the prediction infomation,we design several reinforced matching algorithms.Besides,based on the framework,we propose optimizations to better improve the system performance.From the theoretical analysis and the experimental results,we prove the scheme can efficiently reduce the overall empty distance while protecting users’ privacy.2.Privacy-preserving ride-sharing scheme with minimum detouring route.In most situations,riders will endure a certain detouring time to share a ride with others,but few of existing privacy-preserving ride-sharing schemes can apply matchings for multiple riders while find results with minimum detouring time.To solve this,we propose a privacypreserving ride-sharing mechanism based on path-planning.The mechanism first filters riders on their trips,then compares the detouring time of different rider combinations and routes based on the homomorphic encryption and the garbled circuit technique.Finally,the mechanism outputs the selected rider combination along with minimum detouring route.Through the security analysis and experimental evaluation,we prove the user privacy,the matching accuracy and also the practical efficiency of the mechanism.