Privacy-Preserving Truth Discovery with Task Hiding

Abstract

Mobile crowdsensing has emerged as a popular platform for solving many challenging problems by utilizing users’ wisdom and resources. Due to the diversity of users, the data provided by different users may vary significantly, and thus it is important to analyze user data quality during data aggregation. Truth discovery has been proven to be an effective mechanism to capture data quality and calculate accurate true values via a weighted combination. In spite of the appealing benefits, existing works on truth discovery either fall short of achieving thorough privacy preservation for participating users or cause tremendous computational and communication overhead. In this chapter, we study challenging problems of truth discovery in mobile crowdsensing and present a lightweight privacy-preserving truth discovery scheme, named LPTD, based on the technologies of secure k-nearest neighbor, data perturbation, and matrix decomposition. Through a detailed analysis, we demonstrate that data privacy and task privacy are well preserved during the whole process. Extensive experiments show that our proposed LPTD has practical performance in terms of accuracy, convergence, computational cost, and communication overhead.