Representation and efficient processing of rich qualitative spatial information

Bernhard Nebel, Jochen Renz

Freiburg University, Department of Computer Science

Duration: 1996 - 2002

Project homepage: www.informatik.uni-freiburg.de/~sppraum/.eng.html
 

In this project, we work on the formal foundations of qualitative spatial representations and on the development of efficient inference methods in this area. Starting from existing approaches, we intend to increase the expressiveness of existing formalisms, we work on questions concerning soundness and completeness, and on the determination of computational complexity of inference problems. Based on that, we will develop inference methods, implement them, and evaluate them empirically. The theoretical results will be used in psychological experiments we design and conduct together with the Cognitive Science group MEMOSPACE at the Institut für Informatik und Gesellschaft. Applications of qualitative spatial reasoning can be found in areas such as natural language processing, document analysis, geographic information systems and robot navigation.

Approaches: empirical investigation; modeling; formalization; implementation
Area of Research: artificial intelligence
Topics: spatial reasoning/ inference; qualitative knowledge; navigation
 

Project publications

Publications of project cooperations:

M - MEMOSPACE
H - Landmark Usage
Q - Spatial Inference