Description

Spatial data management is gaining increasing importance, e.g., in the context of geographic information systems, and the marketplace offers standard systems for the management of geographic information. In the research domain, substantial research has been and continues to be conducted in the area of spatial data management. However, little attention has been given to the integration of temporal aspects in spatial data management. By integrating temporal aspects into spatial data management, "closure" is reached: integrated management of all aspects - including temporal, spatial, and other aspects - of data concerning spatially and temporally referenced objects. This is the focus of the present Ph.D. study.

In the Ph.D. study, we more specifically investigate the management of continuously moving objects that may be represented as points in the database. More specifically, the following challenges are addressed.

	A representation for the time-varying position of moving objects is proposed. This representation takes into account the imprecisions caused by measurement errors, sampling, and the representation itself.
	The purpose of indexing is to narrow down the search space when searching for data that qualifies for the answer to a query. The indexing of moving points has only been researched lightly so far and poses new challenges. This Ph.D. study will attempt to meet these challenges by adapting existing indices for the purpose of indexing moving point objects.
	In a variety of moving-point applications, the positions of objects are restricted by what we term infrastructure. People are moving in buildings with floors and walls and thus cannot move freely in three-dimensional space, but are restricted to the building's infrastructure. Similarly, vehicles move in a road networks, trains follow tracks, and ships and airplanes are assigned to predefined routes.

This study will attempt to take such infrastructure into account during query processing, in an attempt to more precisely and concisely represent and predict the positions of moving objects.

A fundamental and very costly operation in databases is the join operation. Its function is to combine data that is somehow related, but is located in different parts of the database.

Moving-point databases are often append-only. Old positions of objects are retained, and new positions pertaining to the present are appended. In this study, we want to reduce the cost of joins on such append-only data with temporal join predicates. Specifically, we propose a sort-merge-based incremental algorithm.

Further readings:

	D. Pfoser and Y. Theodoridis, Generating Semantics-Based Trajectories of Moving Objects [.ps.gz]
	D. Pfoser and C. S. Jensen, Capturing the Uncertainty of Moving-Object Representations [.ps.gz]
	D. Pfoser and C. S. Jensen, Incremental Join of Time-Oriented Data [.ps.gz]
	D. Pfoser and N. Tryfona, Requirements, Definitions and Notations for Spatiotemporal Application Environment [.ps.gz]