Research Of Location Privacy Protection Method Resistant To Semantic Attacks In Road Networks

The advancement of intelligent mobile devices and the progress of science and technology such as GPS,Beidou,intelligent computing,have greatly promoted the development of Location-based Services(LBS).LBS-related applications have emerged in all aspects of daily life and people are more and more inseparable from such services.However,while we enjoy the convenience of personal location information privided to us,it also leads to the exposure of personal location information,even the disclosure of associated information such as personal identity.Currently the frequently used location privacy protection methods are based on K-anonymity and their extension methods,which are structurally strong and easy to implement,However,most existing studies ignored the impact of semantic attacks on anonymous security.To solve above problem,the thesis puts forward corresponding location privacy protection methods for different service scenarios,mainly working as follows:For semantic attacks in a single position service request,a position privacy protection method based on semantic diversity of K-anonymity is proposed.Firstly,in response to the low modeling quality caused by the weight of single in the selection of similar semantic location types to construct an anonymous set by Euclidean Distance,it can establish a semantic sensitivity model under the road network environment constraint,so that the location of different semantics is different.The index mechanism in combination with the differential privacy model is based on semantic sensitivity,user access frequency and other weights to build semantic anonymous sets to improve security in anonymous data during anonymous set construction;Finally,based on the location semantic type of the segment to build anonymous set,the positions of the anonymous set will not be distinguished and the semantic diversity will be met.Through experimental results,the algorithm has a better effect on anonymous success rate and relative anonymity by 5% and 0.5,which can effectively defight against semantic inferior attacks.For semantic attacks during continuous position service request,a location privacy method based on trajectory differential privacy is proposed.Firstly,according to the trajectory of continuous position service request in the road network environment,the disclosure of the user-related information is established in accordance with the leakage problem of the position privacy,and the track semantic privacy attack model under continuous position service query is established;then,in order to resist such an attack,it selects the trajectory similar to the user by cosine similarity to reduce the risk of the anonymous trajectory filtered by the attacker;finally,based on the improved Gaussian mechanism and the construction of the trajectory anonymous set,it counts the access frequency of different semantic types and adds dynamic Gaussian noise,providing more precise release results while improving trajectory privacy,and better taking care of data security and utility.The experimental results show that the algorithm has a good effect on the anonymous success rate,the trajectory similarity is relatively increased by about 0.2,and the relative error and the average error in the noise reduction data are relatively low by 4% and 30,providing more precise position data noise reduction results and it effectively balances the relationship between position data security and availability.