Multiscale classification of moving objects trajectories

In this paper we propose a classification model for moving objects trajectories. We assume that the classification is based on a multiscale map, and we simply define a trajectory pattern as the sequence of zones an object crosses during its travel. These patterns constitute the basis of classification operators. We also define a pattern-based query language which allows an online and continuous classification of moving objects. Finally a prototype which shows the validity of the approach is briefly described.