RPTD: Reliability-enhanced Privacy-preserving Truth Discovery for Mobile Crowdsensing

Abstract

Mobile CrowdSensing (MCS) provides effective data collection through smart devices carried by users. However, the data sensed from various devices is privacy-sensitive and not always trustworthy so that the cloud server needs to extract truthful values while protecting the privacy of user personal data. Although some privacy-preserving truth discovery mechanisms have been proposed to address these issues, they ignore the fact that the reliability of truth discovery algorithms may be considerably degraded by outliers in sensing data, and still cannot guarantee strong privacy. In this article, we propose a Reliability-enhanced Privacy-preserving Truth Discovery scheme (RPTD) for MCS to overcome these shortcomings. First, we design a multi-client inner product functional encryption to fully protect the privacy of sensing data, user weights and inferred truths, while supporting dynamic users. Then a new filtering method is constructed to accurately identify outliers in encrypted sensing data submitted by users, which eliminates the disturbance of outliers on the reliability of truth discovery. Theoretical analysis shows that RPTD ensures practical efficiency in computing and communication overhead while ensuring strong privacy and outliers filtering. Experimental results validate feasibility and effectiveness of the proposed RPTD scheme.