Research of spatio-temporal similarity measure on network constrained trajectory data

Abstract

Similarity measure between trajectories is considered as a pre-processing procedure of trajectory data mining. A lot of shaped-based and time-based methods on trajectory similarity measure have been proposed by researchers recently. However, these methods can not perform very well on constrained trajectories in road network because of the inappropriateness of Euclidean distance. In this paper, we study spatio-temporal similarity measure for trajectories in road network. We partition constrained trajectories on road network into segments by considering both the temporal and spatial properties firstly, then propose a spatio-temporal similarity measure method for trajectory similarity analysis. Experimental results exhibit the performance of the proposed methods and its availability used for trajectory clustering. © 2010 Springer-Verlag Berlin Heidelberg.