Extracting the Main Routes and Speed Profiles Between Two Locations from Massive Uncertain Historical Trajectories

Abstract

Increasing penetration of intelligent Internet of Things (IoT) makes it possible to track locations in real time. Capturing the spatiotemporal patterns of a group of mobile objects from their GPS trajectories is a challenging task but crucial to many related applications. This study investigated the trajectories of mobile objects between two locations and developed a methodology to figure out the main routes and their associated speed profiles from massive historical trajectories. DTW (Dynamic Time Warping) and HDBSCAN (Hierarchical Density-Based Spatial Clustering of Applications with Noise) were employed to calculate the similarity of trip trajectories and automatically cluster trajectories. A break-merge approach was employed to develop the main route and speed profile of each cluster which could be further used to reconstruct trajectories with big uncertainty. Experiments on the real AIS (Automatic Identification Systems) trajectories of container ships between the terminals of Shanghai and Ningbo port indicates the effectiveness of this work.