What Logical Model Is Suitable for Relational Trajectory Data Warehouses? Application to Agricultural Autonomous Robots

The ability to query vast amount of historical data for statistical analysis and reporting is provided by Data Warehouses. They facilitate Business Intelligence for effective decision-making significantly. In recent years, great progress has been made in movement monitoring devices, such as smart phones and GPSs. The storing and managing of spatio-temporal data related to the trajectories of moving objects in a data warehouse is called Trajectory Data Warehouse (TDW). The relational approach is adopted widely for the logical representation of TDWs, since it is based on the classic database approach where data representation and processing are handled on structured data. In this paper, the key idea is to consider different logical relational TDW models, i.e. flat, segment and complex, which are compared and evaluated. The study is based on a novel classification of OLAP queries, the cardinality of facts and the resolution of each trajectory in segments. Real data provided by agricultural autonomous robots is used, where experiments on size and time performances are conducted and discussed.