Research On Moving Object Privacy Protection Echnology In Location Based Service

In recent years with booming development in wireless communication and mobilepositioning technologies, application of location based services (LBS) have been gainingincreasingly popularity. On the one hand, LBS bring people convenience services. On theother hand, the privacy problem has received widespread attention. The researchers putforward many solutions on the problem, but in actual road network environment they maymake trouble that calculated distance is not accurate or user’s location is not reasonable andso on. It is a very important research direction how to effectively prevent the attacker fromobtaining the user’s location and identity information and determine who submit servicerequest in LBS privacy preservation.Our study is base on the problem mentioned above, and proposes a new type of locationprivacy protection technology, which involves the user’s location or identity, consideringcontinuous query in the actual road network environment. Details are as follows:Firstly, the clustering method is adopted to analysis and discuss the distribution ofhistorical user. In order to ensure the credibility of generated dummy user and cut off thecorrelation of two consecutive pseudonyms, we analyze the distribution of historical users. Atfirst by using based on density DBSCAN clustering algorithm to gain clusters of arbitraryshape, that is the densest area in which the distribution of historical user. And then, accordingto different privacy demand, the users choose to adopt dummy generation or pseudonymchange method for privacy preservation. The part of research is basic and pretreatment as thefollowing two methods.Secondly, in term of the user’s location information, we proposes a kind of based onhistorical users distribution Dummy Generation method (BHUDG). In BHUDG method, weobtain the dummy user’s initial position based on historical users’ distribution for staticprivacy protection, in the user continuous query, dummy user with real users keep speedconsistency. It can enhance the dummy user to intend to truth.Thirdly, for user’s identity information, we propose a kind of based on historicalusers’distribution the pseudonym change method (BHUPC). In LBS privacy protection, thepseudonyms can help to hide the true identity of the user, and changing pseudonyms in themix-zone can break the correlation two consecutive pseudonyms, so the core issue of the pseudonym change methods is to find a simple and effective mix-zone. We analysis thedistribution of historical user to the densest areas, and identify them as mix-zone, then torealize the purpose of privacy protection. Under normal circumstances, the number of movingobjects will be relatively more at road intersection, so also it as mix-zone.Finally, we verify the above two based on the historical users’ distribution privacyprotection method is correct and effective, through experiments under different criteria.