The subject of this thesis, qualitative spatial reasoning, falls between two areas of artificial intelligence research: qualitative reasoning, and robotics. The work described draws on past research in both of these fields, and has resulted in the development of a new technique for qualitative representation of shape and space, together with methods for performing simple spatial reasoning tasks using the representation. I intend that the results presented here should be useful in finding new approaches both to robot planning and problem solving, and to qualitative reasoning, while being of interest to a more general audience insofar as they represent a plausible computer model for certain aspects of human spatial reasoning.