An Attribute Generalization Mix-Zone Without Privacy Leakage

In the road network, mix-zone is usually considered as an efficient application that provides a balance between privacy protection and service quality. At the same time, the mix-zone can also resist the attack of tracking without any affection in the quality of feedback result of navigation. However, just as the old adage goes, nobody is perfect, the mix-zone also has two non-negligible deficiencies. The first one is that the mix-zone is difficult to cloak or generalize all types of the attributes emitted by the user. The other one is that the mix-zone is difficult to resist the attack initials by the disguiser who participates in the process of constructing the anonymous group. In order to cope with above-mentioned two problems, this paper provides an uncorrelated mix-zone based on the conception of attribute generalization and homomorphic encryption. In this algorithm, the process of attribute generalization is executed by an agency who is the winner in the private bidding, and the value of generalized attributes value is calculated by the secure multi-party computation under the principle of homomorphic encryption. Accordingly, the private information of the user in the mix-zone is preserved and any entities, no matter the agency or other participants, are prevented from any information about the user as there is no information been exposed. Furthermore, as the attribute of each user is generalized, for the mix-zone leavers, their attributes emitted are similar to each other; it will be difficult for the adversary to correlate any user with the attribute grasped before the user enter the mix-zone. Therefore, the potential hazard of adversary tracking the user with attribute comparison is solved. At last, in order to demonstrate the performance of our proposed algorithm in theoretical and practical fields, several analyses of security and simulation experiments are proposed, and the verification of security as well as the results of simulations and comparisons can further demonstrate the priority.