On unified spatiotemporal data model
Peiquan Jin,Lihua Yue,Yuchang Gong
2005-01-01
Abstract:Spatiotemporal data model, which is the basis of spatiotemporal data management, deals with the representation and manipulation of spatiotemporal data. So far many spatiotemporal data models have been proposed, but most of them are towards specific applications and few are for general purposes. It is not feasible to design respective data management systems (DBMS) for different applications because it is expensive and exhausting. Aiming to provide general support for spatiotemporal applications in DBMS, our idea is to design a unified spatiotemporal data model and then to implement a general spatiotemporal DBMS. This idea is much like the development of relational database technology. Traditional relational DBMS can support general applications, and is not for specific applications, and most importantly it is successful in real applications. In this paper, we study several issues on the unified spatiotemporal data model, including the data structure, the querying operations and the implementation issues. The unified spatiotemporal data model is based on the object-relational data model. Spatiotemporal objects are represented as spatiotemporal relations and spatiotemporal tuples. And queries on spatiotemporal relations are implemented through an extended relational algebra. The spatiotemporal changes of spatiotemporal objects are represented by spatiotemporal data types, and queries on spatiotemporal data and changes are supported by the operations defined on those data types. The data structure of the unified spatiotemporal data model, the querying operations along with the algorithm to transform the data model to object-relational database management systems are discussed in detail. The unified spatiotemporal data model has been partly implemented on Informix, and a real example about China Historical Geographic Information System (CHGIS) is discussed. The experimental results show that the unified spatiotemporal data model is able to represent, store and query spatiotemporal changes successfully. * Corresponding author.