Research On Location Privacy Protection Based On K-nearest Neighbor Query In Internet Of Vehicles

Vehicle self-organising networks（VANETs）can improve traffic flow,promote intelligent mobility and provide convenient information services,with a wide range of applications such as assisted vehicle driving and safety warnings.But as data sharing between vehicle nodes and other nodes explodes in scale and dimensionality,traditional VANETs can suffer from issues such as identity validity and message reliability.At the same time,existing scalable incremental processing techniques for nearest neighbour query processing only provide technological innovations on tree indexing without considering hierarchical privacy protection issues,resulting in inefficient queries and low query accuracy.For these reasons,this paper investigates continuous k-nearest neighbour query algorithms in connected vehicle systems.This paper firstly investigates the data secure sharing and storage scheme based on full homomorphic encryption and federated blockchain（DSS-AHCB）,and proposes for the first time a scheme for secure data encryption,decryption and transmission based on full homomorphic and federated chain.The digital signature technique based on the bilinear pair nature of elliptic curves ensures the reliability and integrity of data transmission to nodes,and introduces the emerging federated blockchain technology to provide a decentralised and reliable database for data storage;to achieve better security performance,a VP~2-AES-FHE encryption scheme is proposed with low overhead,plaintext extension and low complexity;in the DSS-AHCB network system The trigger conditions for pre-selected nodes are restricted through smart contracts when transmitting and storing data,and data coins are allocated to vehicles involved in data contribution.Theoretical studies and experiments show that the proposed DSS-AHCB solution improves the efficiency of encrypted data transmission and is more secure and reliable in data sharing and storage.A continuous k-nearest neighbour query algorithm based on privacy hierarchy and combined similar sets is further investigated.Based on the TPR tree index,pruning is performed according to three pruning rules;at the same time,the combined similarity set is used to delineate the pre-search area to calculate query results,which improves the availability of data based on the hierarchical differentiation of privacy protection levels;on the basis of considering the complex association between sensitive and non-sensitive data attributes,the Bayesian network is used and the associated information is calculated in the case of publishing multidimensional related high-dimensional data values to improve the accuracy of queries.Theoretical studies and experiments show that the proposed algorithm improves the efficiency of the algorithm under multiple privacy requirements and makes the query results more accurate for consecutive k-nearest neighbours.