Academic journal article Cartography and Geographic Information Science

Representation and Manipulation of Moving Points: An Extended Data Model for Location Estimation

Academic journal article Cartography and Geographic Information Science

Representation and Manipulation of Moving Points: An Extended Data Model for Location Estimation

Article excerpt

Introduction

Geographic information--commonly represented by points, lines, and polygons and their locations--is generally stored on proprietary GIS systems. These systems are, however, less effective than traditional Database Management Systems (DBMS) when it comes to non-dimensional data, often referred to as attribute data. Hence, some hybrid systems combine GIS features, thereby taking advantage of their efficient spatial data access and graphic display capabilities, with traditional DBMS to handle attribute data. This approach typically suffers from poor integration of the two architectures and difficulties in maintaining the integrity of spatial and attribute data. Spatial databases are meant to overcome this problem by storing all data, attribute and spatial, within a single data model.

None of these systems considers another natural dimension, the temporal one. Geographic information is subject to change over time in two relevant aspects: the geometry or location of the represented objects, and the associated attributes. We focus in this paper on representing the evolution of the spatial features of objects. This contrasts with the main focus of research on temporal database systems which focus on the modeling, representation, and querying of temporal attribute data (Soo 1991; Tsotras and Kumar 1996).

Despite the many applications of spatio-temporal systems there are only a few approaches dealing with the integration of spatial and temporal components. The features and requirements of such systems are largely heterogeneous. For example, some systems handle objects with continuously changing properties while others handle objects with properties that change at discrete steps. The properties that change differ from one system to the other.

Research on the representation of evolving data has so far focused on the study of discontinuous change, recording either isolated events or changes of state. With isolated events, the attribute values are valid only for a time instant. In the latter case, attribute values are valid and constant until an event produces a change of state. In contrast, continuously changing data are characterized by potentially having a different value at each instant. This particularity strongly influences their management.

In this paper, we introduce a classification of spatio-temporal applications based on the properties of the represented objects. A key distinction is made between continuous and discontinuous change. We specify a data model for the representation of mobile objects that can be modeled by points. Because it is not possible to store exact knowledge on the movement of objects, i.e., all their possible states, we propose a semantics to cope with the uncertainty of the position of objects. We then analyze certain functional aspects of the implementation of the data model in a Relational Database Management System (RDBMS).

The rest of this paper is organized as follows. The following section introduces a classification of spatio-temporal systems based on the properties of the represented objects. We then present an overview of relevant work in the domains of temporal and spatio-temporal systems. We do not include purely spatial systems in this overview, as the main point is to discuss moving points for which spatial size and shape are irrelevant. An excellent survey may be found in Gutting (1994). The following section presents a data model for the representation of moving objects and an extended semantics to cope with imprecise answers in query operations. We then discuss the implementation of the data model in a RDBMS, pointing out some difficulties and suggesting further developments. The final section concludes and presents directions for future work.

Classification of Spatio-Temporal Applications

A search for existing and potential applications of spatio-temporal systems shows that the application domains are vast and that systems may include objects that have greatly differing properties. …

Search by... Author
Show... All Results Primary Sources Peer-reviewed

Oops!

An unknown error has occurred. Please click the button below to reload the page. If the problem persists, please try again in a little while.