An optimal privacy-preserving mechanism for crowdsourced traffic monitoring
Yunhua He, Limin Sun, Zhi Li, Hong Li, Xiuzhen Cheng · 2014
Crowdsourced traffic monitoring employs ubiquitous smartphone users to upload their GPS samples for traffic estimation and prediction. The accuracy of traffic estimation and prediction depends on the number of uploaded samples; but more samples from a user increases the probability of the user being tracked or identified, which raises a significant privacy concern. In this paper, we propose a privacy-preserving upload mechanism that can meet users\textquoteright~diverse privacy requirements while guaranteeing the traffic estimation quality. In this mechanism, the user upload decision process is formalized as a mutual objective optimization problem (user location privacy and traffic service quality) based on an incomplete information game model, in which each player can autonomously decide whether to upload or not to balance the live traffic service quality and its own location privacy for utility maximization. We theoretically prove the incentive compatibility of our proposed mechanism, which can motivate users to follow the game rules. The effectiveness of the proposed mechanism is verified by a simulation study based on real world traffic data.