Vehicle Trajectory Obfuscation and Detection

A vehicle in road networks shares location data with other vehicles and location-based services (LBS) through Internet-of-vehicles (IoV). By analyzing the location data from vehicles, LBS providers can offer vehicles better services. However, fake trajectories created by adversaries and malicious drivers diminish the location data utility in IoV and breach the quality of LBS. Illegal trajectory detection is vital to ensure location data utility in IoV. Existing location privacy-preserving schemes like obfuscation schemes add noise to actual location data increasing difficulties in detecting illegal trajectories. In this paper, we detect illegal trajectory in the case that all drivers in road networks protect location privacy by using obfuscation. We propose a new personalized obfuscation mechanism to dynamically and adaptively protect the location privacy of drivers in road networks. Considering uneven protection, we propose a trajectory detection scheme to classify trajectories in IoV. We evaluate our detection method with the data of real-world road networks, which is an important scenario of smart cities. The experiment results show that the proposed classifier outperforms existing studies in detecting malicious obfuscated trajectories with at least 94% of the Area Under the Curve (AUC) score.