Logistics Management in a Mobile Environment: A Decision Support System Based on Trajectory Mining

Location prediction systems have become of interest in several application domains involving ubiquitous systems in the recent years. We propose a novel approach to trajectory prediction, that supports logistics optimization and decision making in a service area of a container terminal at a freight seaport. We devise a prototypical wireless infrastructure which collects historical trajectory data from the movements of containers. The approach consists in the application of a data mining technique for the development of the underlying location-aware prediction model that supports multiple representations and granularities both in space and in time. Also, we devise a path-prediction technique, capable of effectively taking into account inherent peculiarities of moving objects, such as a varying number of trajectories followed by each individual object and possible differences in the length of such trajectories.