PPTD: A Privacy-Preserving Truth Discovery Scheme in Mobile Crowdsensing

Truth discovery (TD) can discover ground truths of sensing tasks from the submitted reliable or unreliable sensing data. However, the existing TD algorithms mainly focus on the privacy preservation of mobile users, and few studies consider the reliability of mobile users when they participate in calculating the ground truths of sensing tasks. To simultaneously ensure the reliability and privacy preservation of mobile users, this paper proposes a privacy-preserving truth discovery scheme in mobile crowdsensing. Firstly, we select reputation value to evaluate the reliability of mobile users. Meanwhile, to preserve the privacy of reputation value, we define two novel concepts, namely reliability level and contribution degree, to represent the reliability and contribution of each mobile user in calculating the ground truths of sensing tasks. Secondly, according to the contribution degrees, we redesign the process of truth discovery which can publicly verify the contribution degrees of mobile users and improve the accuracy of ground truths of sensing tasks. Finally, theoretical analysis shows that our scheme is private and secure. The comparisons with the existing schemes show that the error rate, mean absolute error, and root mean squared error in the proposed scheme are reduced by about 54%-63%, 13%-34%, and 15%-46%, respectively.