VCLT: An accurate trajectory tracking attack based on crowdsourcing in VANETs

Abstract

We investigate trajectory tracking in Vehicular Ad hoc Networks (VANETs) in this work. Previous tracking methods suffer from low accuracy, large overhead, and big error. In this paper, we propose a Vehicular Crowdsourcing Localization and Tracking (VCLT) scheme for mounting a trajectory tracking attack. In our scheme, crowdsourcing technique is applied to sample the location information of certain users. Then matrix completion algorithm is used to generate our predictions of the users’ trajectories. To alleviate the error disturbance of the recovered location data, Kalman filter technique is implemented and the trajectories of certain users are recovered with accuracy. At last, extensive simulations are conducted to show the performance of our scheme. Simulations results reveal that the proposed approach is able to accurately track the trajectories of certain users.