Design and Implementation of Spatial-temporal Data Model in Vehicle Monitor System

In recent years, Global Positioning System (GPS) has been greatly utilized in collecting real-time geographic data for a mobile unit. With the latest developments in the technologies of Geographic Information System (GIS) and wireless communications, the location data collected by a GPS unit could be easily displayed in a real-time fashion. However, for large scale applications such as a taxi fleet, the task to store and query the database with continuous GPS location data collected over time still remains a challenge. In this paper, a spatial-temporal model that is designed specifically for managing extensive amounts of GPS position data is described. The model consists of two major components, i.e., cube cell and trajectory. A cube cell is a three-dimensional space determined by a pre-defined time interval as well as the x and y axles for spatial coordinates. A trajectory, on the other hand, has a time attribute and may be stored as a line object in software applications such as Oracle Spatial. In the model, the position data that form a moving trajectory is stored in each cube cell to record a given object's locations over the specified time interval and a trajectory is defined as the basic visiting unit. The data structure allows extensive GPS data to be efficiently stored and queried. The model is also capable of eliminating redundant data for objects remaining at the same locations, which may further reduce the storage requirement. For evaluation purpose, the model was implemented in a Microsoft Windows environment using the GPS data collected from the taxi monitoring system in Guangzhou, China. Oracle 9i and Oracle Spatial were used in the model to create and query the database. The model was shown to significantly reduce storage requirement and greatly enhance the query speed.