Supervisor

Here the dedication The development of new technologies for mobile devices and low-cost sensors re-sults in the possibility of storing large data volumes about trajectories of moving objects. In this thesis we propose a multi-dimensional data model to store aggre-gate measures computed over such huge data volumes. This allowed us to define a Trajectory Data Warehouse (TDW) that is loaded by managing and transform-ing a data stream of spatio-temporal observations of moving objects, arriving in a irregular and unbounded way. The stream nature of our huge input data, concerning moving object observa-tions, may make it difficult to build and maintain a data warehouse. The trans-formation and loading phase of our TDW is particularly challenge, since it has to work under resource constraints (memory, processing), by accommodating possible irregular arrival rate of input data. In our TDW the identifier of the trajectories are abstracted in favour of aggregate information concerning global properties of