A fundamental step in the development of an AI system (or any computer system) is designing appropriate computer representations for entities in the subject domain, for the properties of those entities, and for relationships between those entities. (The entities may be physical objects, computer models of objects, or abstract facts). The resulting representation then has a profound effect on the structure, operation, and capabilities of the whole system. This is particularly noticeable in the field of qualitative physics, because the distinctive nature of the field is largely in the novel representation used, and a completely new approach to computer reasoning has developed as a result of that representation.
The application of qualitative reasoning techniques in a spatial domain depends primarily on the development of a (qualitative) representation for physical objects and spatial relationships between objects.
Techniques for the representation of physical objects using computer systems have been developed in the context of several different fields. This chapter surveys the field of robotics, pointing out the approaches that have been taken to spatial representation and reasoning in the past. For comparison, the final section of the chapter briefly surveys shape representation methods used in the areas of computer graphics, and computer aided design. Some of the representation methods that have been used in these fields can be adapted to make a basis for qualitative representations, and this will be discussed further in chapter 4.