Native Trajectory Database Based On PostgreSQL/PostGIS

As the prevalence of the application of mobile positioning facilities, people are more and more relying on the positioning facilities. For example, one will use the position system to deal with the travel planning, or when one gets lost in strange neighborhood, position system must be helpful, and also, people will addicted to location based social games with GPS systems. Trajectory data can be of significant use in many areas of research, such as research on traffic flow, animal migration, people behavior research and so on. Therefore, one of the hot issues in the realm of trajectory database research is effective collection, well-management, query and analysis of huge amount of trajectory data.This article thoroughly analyses and compares current solutions to manage trajectory data, and designs a kind of trajectory data model which is based of relation-object based database system and the idea of using user extension. Design of Trajectory data model includes devising logical data model and physical data model. Logical data model is designed for construction of relation between objects. And physical data model is aiming at how the objects stored in physical disks. Furthermore, spatial and temporal query interfaces are designed for basic use of trajectory data, such as queries between trajectories and points, queries between trajectories and regions as well as queries between trajectories. Indexes are also built to improve query speed.This research has put the design of the trajectory data model into implement as native database in PostgreSQL, including self-definition database and extension query interfaces. And series of tests for evaluate the performance of the trajectory database has been taken. Results of these tests show that the trajectory data model can improve the efficiency of space of storage substantially, and has been convenience for users to utilize spatial-temporal queries. Thus, the model is pragmatic.