Efficient Multilevel Threshold Changeable Homomorphic Data Encapsulation With Application to Privacy-Preserving Vehicle Positioning

Tianhui Zhou, Jun Zhou, Zhenfu Cao, Xiaolei Dong, Kim-Kwang Raymond Choo

Abstract: Although the global navigation satellite system (GNSS) has been successfully applied in search and rescue operations for locating lost or damaged vehicles due to its significance in precise positioning, there are still challenges. To enhance the accuracy of localization, positioning can be done with the collaborative estimation provided by neighbouring mobile terminals as reference vehicles. Meanwhile, there are security and privacy implications associated with such an approach – e.g., potential for privacy leakage of both the positioning-related data (e.g., positions of reference vehicles, distances between reference vehicles and the target vehicle) and the estimated positions of the target vehicle. Such concerns are important to address in deployments for sensitive applications such as defense. For example, the location information of a damage vehicle on the battlefield should only be securely evaluated by search unit and be successfully decrypted by an authorized set of officers in rescue unit with the required authorization levels. Accordingly, the threshold should be flexibly allocated and changed for types of security surroundings. However, existing techniques of threshold public key homomorphic encryption approaches are not only computationally and communication intensive, but merely support a fixed pre-defined threshold. To address these challenges, we propose an efficient multilevel threshold changeable homomorphic data encapsulation mechanism (MCTh-HDEM). In MCTh-HDEM, we leverage the technique of multilevel threshold changeable secret sharing in order to support both batch encryption and lightweight matrix calculations in the encrypted domain, and also multilevel threshold changeable decryption. Then, we design a lightweight privacy-preserving vehicle positioning scheme (PPVPS), by refining our proposed MCTh-HDEM. The position of the lost and damaged target vehicle on the battlefield would be efficiently inferred by a set of reference vehicles in search unit while protecting positioning related data, and the target vehicle location can be flexibly decrypted by rescue unit. Finally, we give the formal security proofs of our proposed MCTh-HDEM and PPVPS. The performance evaluation and extensive experimental results demonstrate the efficiency and accuracy of our proposal.