EPNS: Efficient Privacy Preserving Intelligent Traffic Navigation from Multiparty Delegated Computation in Cloud-Assisted VANETs
Jun Zhou, Shiying Chen, Kim‐Kwang Raymond Choo, Zhenfu Cao, Xiaolei Dong
Abstract
Real-time navigation is a fundamental service with the emerging techniques of intelligent transportation and crowdsensing. Unfortunately, the breach of location privacy significantly impedes its wide adoption. Most existing state-of-the-art exploit either pseudonyms or public key fully homomorphic encryption (FHE) to protect location privacy, which requires an online certificate authority (CA) or loads intolerably heavy computational/communication overhead on resource-constrained vehicular users. In this paper, a new cryptographic primitive named efficient multiparty delegated computation (MPDC) is firstly proposed, where any one-way trapdoor permutation is required to perform constant times (i.e. twice) on each resource-constrained data provider, to encrypt a batch of messages and enable types of secure evaluations over ciphertexts encrypted under multiple keys of multiple parties. Based on MPDC, we devise a lightweight privacy-preserving real-time intelligent traffic navigation scheme (EPNS) in cloud-assisted VANETs. The proposed approach predicts an optimal driving route of shortest time without disclosing either vehicular users' private location or the navigation result to the collusion between the semi-trusted cloud/cryptography service provider (CSP) and unauthorized users, by securely evaluating auto-regression moving average (ARMA) model with spatiotemporal correlations of high accuracy and efficiency. Finally, formal security proofs and the extensive evaluations demonstrate the utility and the practicability of our approaches.