Dual-side privacy-preserving task matching for spatial crowdsourcing

Abstract

With the popularity of mobile phones and the ubiquity of wireless transmission technologies, spatial crowdsourcing (SC) has emerged as a novel approach to outsource location-based tasks to a set of workers who physically move to the designated locations to perform the tasks. To achieve the accurate task matching, both requesters and workers need to expose their locations or queries to the SC-Server, which raises security concerns. Although many protection measures have been proposed, there are some drawbacks in one-side protection, dual-server setting and user scalability when they are applied to the practical crowdsourcing environment. In this paper, we design a general framework for spatial task matching in a single-server setting to simultaneously protect the privacy for both tasks and workers. Combining multi-user searchable encryption with segment tree, we propose two different schemes to achieve the spatial task matching over the encrypted data. Efficient user enrollment and revocation are also supported. Extensive experiments validate the feasibility of our schemes.