Research On Lightweight Privacy Preserving For Location-based Services In Wireless Vehicular Networks

With the emergence of intelligent transportation and crowd-sensing techniques,the self-organized wireless Vehicle Ad-Hoc Network(VANET)and its applications have received widespread attention.VANET serves the fundamental basis of location-based service(LBS),which collects and distributes different types of data among vehicles,such as traffic conditions and shopping information,and provides location based services,such as navigation,social networking and precise advertising recommendation.Security and privacy issues including data confidentiality and message authentication are critical to the wide application of LBS.Unfortunately,there still exists a series of unaddressed challenging issues in the existing works:1)The most commonly used technique,pseudo identity,can protect the privacy of vehicular real identity,but requires frequent interaction with the Trusted Authority(TA)to obtain anonymous certifications or pseudonyms.And it would be vulnerable to background attacks,due to the fact that LBS message is not directly hidden by encryption.2)Resource-constrained on-board units tend to bear high computation/communication overhead,usually for message encryption/decryption or authentication;3)Adjacent vehicles often generate a large number of redundant messages,which will heavily waste system resources.In order to solve the above problems,the main contributions of this dissertation include the following two parts:First,we devise a lightweight privacy-preserving outsourcing traffic navigation scheme in cloud-assisted VANET(EPNS),in which LBS functions are outsourced to cloud servers to reduce the overhead of vehicle users,while utilizing secure outsourced computation technique to protect user's locayion privacy.In this dissertation,a new cryptographic primitive named efficient multiparty delegated computation(MPDC)is proposed as a building block.In MPDC,resource-constrained data senders can encrypt multiple messages in a batch manner by performing any one-way trapdoor permutation just constant times,and we design several privacyprotected basic arithmetic protocols over the multi-key ciphertext field,including addition,subtraction,multiplication,and less than or equal operations.Then based on MPDC,we designed EPNS.In EPNS,a real time short-term traffic prediction model with high precision is outsourced to the cloud server,and is used to predict future road conditions.Then cloud server returns to the request user an optimal driving route with shortest driving time without disclosing either vehicular users' private location information or the navigation result to the collusion between the semi-trusted cloud/crypto service provider(CSP)and unauthorized users.Finally,the security of proposed MPDC and EPNS are formally proved,and the extensive evaluations demonstrate the utility and the practicability of our approaches.Second,in distributed LBS without cloud server,we propose an efficient privacypreserving distributed LBS bundle authentication scheme(SAVE)through secure redundancy filtering in self-organized vehicular social networking.Initially,a privacypreserving message filtering mechanism with dynamic vehicle group formulation is designed from the message sender's aspect by exploiting lightweight one-way hash chains.It eliminates duplicate information from aggregation to LBS bundles and dramatically reduces both communication and computational complexity for authentication.Especially,an online/offline aggregate signature is improved to achieve batch verification to further improve the efficiency of LBS bundle authentication,based on any one-way function holding the property of multiplicative homomorphism.Then,to achieve efficient vehicle group key updating and resist free-riding attack without the intervene with TA,an improved self-healing key distribution with distributed revocation is proposed.The identity privacy and two levels of location privacy are well protected.Finally,the formal security proof and extensive simulation results verify the practicability and efficiency of our proposed SAVE.