A Semantic-Based Data Model for the Manipulation of Trajectories: Application to Urban Transportation

Nowadays, the increasing development of positioning and wireless communication technologies favors a better real-time integration and manipulation of large spatial databases. This offers many new opportunities for the development of trajectory databases, but a number of research challenges are still open as the generated information is often unstructured, continuous, large and sometimes unpredictable. The research presented in this paper develops a modeling approach that integrates the semantic, spatial and temporal dimensions when representing spatial trajectories at the abstract and logical levels. A data manipulation language that supports the querying and analysis of large trajectory databases is also proposed. The spatial database model is based on algebraic data types, and a prototype is developed on top of the DBMS PostgreSQL/PostGIS. The whole approach and the prototype development have been experimented and applied to benchmark transportation data derived from an origin-destination survey in the region of Quebec in Canada.