Spatial reasoning for robots: a qualitative approach

Abstract

This thesis presents a new approach to the problem of shape representation for high level robot reasoning tasks. Techniques from the field of qualitative physics are combined with current methods for solid modelling to develop a qualitative representation of two dimensional shape and position. The qualitative spatial representation has been used to solve simple spatial reasoning tasks. The results of this investigation are applicable both to the field of robotics, where they provide a new approach to programming and control, and to the field of qualitative physics, which has been hampered by a lack of general purpose spatial reasoning techniques.