Spatiotemporal-Aware Privacy-Preserving Task Matching in Mobile Crowdsensing

Tao Peng, Wentao Zhong, Guojun Wang, Shaobo Zhang, Entao Luo, Tian Wang

Abstract: Task matching is widely used for participant selection in mobile crowdsensing (MCS). However, accurate task matching relies on collecting a large amount of user information, which has the risk of privacy leakage. Existing privacy-preserving task matching methods have the disadvantages of low matching efficiency and coarse matching granularity, and are difficult to apply to MCS because of higher real-time requirements. In this article, we propose a spatiotemporal-aware privacy-preserving task matching scheme, achieving efficient and fine-grained matching while protecting privacy between users and task publishers. Specifically, the time matching score (TMS) and location matching score (LMS) between users and tasks are defined for the spatiotemporal requirement of MCS. In addition, a lightweight protocol called SCP (secure computing protocol) is constructed based on Shamir secret sharing and Carmichael theorem for securely calculating TMS and LMS and matching attribute values by size and range. The correctness and security of our scheme are proved by detailed theoretical analysis, and the experimental result shows that the computational overhead of our proposed scheme is only 10% of that in the scheme we compared with, while the difference in communication overhead is less than 200 KB.