Context-aware similarity of GPS trajectories

Abstract

Measuring similarity of GPS trajectories has attracted a lot of attention in recent years. As a result, multiple trajectory similarity measures have been developed and are used in a wide set of applications which aim to extract meaningful information from large collections. In this paper, we focus on some of the most popular measures and study how they all can be adapted to use contextual information. We experiment using the buildings in an urban setting as the context and demonstrate how it impacts the similarity values. Experiments show that routes rank differently in terms of similarity in the presence of context which can have serious implications in applications such as trajectory search and clustering similar trajectories.