Discovering Corridors From GPS Trajectories

Abstract

The increasing pervasiveness of GPS-enabled devices results in the collection of massive trajectories datasets. The vast amount of the generated location data is particularly difficult to be processed, interpreted and analyzed, due to its complexity. Nevertheless, in many cases a considerable number of moving objects share common paths and their whole trajectory can be decomposed as a sequence of such commonly accessed paths, referred as corridors. In this paper we formulate the problem of corridor discovery using GPS data that represent user trajectories. We initiate research for developing an algorithm to solve this problem efficiently and we present initial experimental results that demonstrate our approach.