SPACENET -- Background

What is SPACENET?

SPACENET is an EU funded Human Capital Mobility Network that brings together the eleven major European spatial reasoning groups. Participating laboratories are sited in eight European countries, including both EU and EFTA states. The multi-disciplinary nature of spatial reasoning is reflected in the inclusion of Computer Science, Artificial Intelligence, Cognitive Science, Linguistics and Geographic Information System groups.

What are the objectives of SPACENET?

Integrate existing European research on qualitative spatial reasoning.
Provide a framework for bilateral and multilateral communication between researchers thus supporting European research in this area.
Provide mechanisms for training and enhancing the skills of workers in this field of research.
Provide means of dissemination of research from within the network both internally and externally to the network.

How will these objectives be met?

A series of SPACENET workshops will bring together scientists from all partner sites, together with a number of researchers outside the network, on a regular basis.
A program of bilateral exchange visits between partners will seek to foster co-operative research projects.
Workshop reports and an electronic news network will enable partners to share research information.
The network is actively engaged in producing a comprehensive textbook that will cover all aspects of qualitative spatial reasoning.

What is qualitative spatial reasoning?

Qualitative spatial reasoning is a field which has defined itself over the last few years as researchers in a variety of subject areas have recognised the extent to which they have interests in common.

In all these areas, sophisticated automated reasoning about the spatial relations between physical objects or regions of space is of fundamental importance; and in many cases, this must be done without precise, quantitative information about these relations. Typically, some knowledge of the topological relationships between the entities of interest may be available, along with incomplete and imprecise information about distances, directions and relative sizes; and from this partial information, useful conclusions must be drawn. Examples of the kind of question for which qualitative spatial reasoning is required can be drawn from the fields mentioned: Identify the islands in the lake and the largest one. Which parts of the network of tunnels can the robot traverse without getting stuck? Could the collection of objects in the scene fit together to make a spherical or cylindrical shell? When cog A is turned clockwise, will cog B turn and if so, in which direction?

These are all examples of the kind of problem human beings solve (and sometimes fail to solve) without making precise measurements; if we are to maximise the potential of computer systems to help them, we must understand the principles that make possible these forms of reasoning. This is not to assume that computer systems will necessarily use the same methods as human reasoners; but the fact that people can answer such questions constitutes a form of proof that usable methods exist. The qualitative spatial reasoning community has set itself the task of finding them.