An efficient and privacy-preserving truth discovery scheme in crowdsensing applications

Abstract

Truth discovery is a reliable and effective technique to resolve conflicts of heterogeneous data and estimate user reliability in mobile crowdsensing systems. Despite its effectiveness, the widespread adoption of truth discovery requires solid privacy preservation against users’ sensory data and reliability information. Existing works of private truth discovery are primarily based on conventional cryptographic primitives, which introduce tremendous workloads on the system. In this work, we first propose an efficient and privacy-preserving truth discovery framework (EPTD-I) by adopting a novel data perturbation mechanism. EPTD-I not only protects users’ privacy but also introduces little overhead on the user side. Moreover, for high mobility environments, we improve the design with a user non-interactive scheme named EPTD-II to shift all encrypted truth discovery operations to cloud platforms. In EPTD-II, each user’s sensitive information is also kept private during the complete truth discovery procedure. Thorough security analysis demonstrates that our proposed schemes are secure and offer a high level of privacy preservation. Extensive experiments conducted on practical and simulated crowdsensing applications demonstrate the effectiveness and efficiency of the proposed schemes.