A density-based algorithm for detecting anomalous trajectories

Fast growth of mobile sensing technologies, like GPS in smart phones, made capturing position data in the form of trajectories easy. But detecting anomalous trajectories, which are grossly different from remaining trajectories, is a major challenge in the surveillance domain and a big data problem. In the present work, a novel density-based method has been proposed and implemented for detection of anomalous trajectories. We introduced the concept of Hausdorff-neighborhood-trajectory, which makes our approach capable of handling border trajectory situations.