next up previous contents
Next: Representation of Detail Up: Two Methods for Qualitative Previous: Two Methods for Qualitative

Representation Issues in Qualitative Robot Reasoning

Comparing the way that humans think about physical situations to the way that robots are currently programmed suggests a number of important issues that should be considered in the development of new robot reasoning systems. In this section, I concentrate on the issues arising from four aspects of human spatial reasoning that appear to be particularly important in solving simple qualitative spatial reasoning problems. These aspects are:

1.
Representation of detail at multiple levels. People are able to store a large amount of detailed information about a complex object, yet also consider that object in terms of gross shape alone when this is necessary. They are also able to focus on a particular detail of the overall shape, while retaining a record of its context. An example of this ability is the way that a mechanic views a car. He knows a huge amount about its detailed shape, but is able to think, when necessary, simply in terms of its overall shape (when driving it).
2.
Independent reasoning in local contexts. Where overall shape is very complex, people are able to reason about one part of the overall shape, treating it as an independent context. The car mechanic for example, when fixing a handbrake, is able to work purely with that local context within the car. Other contexts, such as the seats or headlights, can be temporarily forgotten. He can also fix a truck handbrake as easily as a car handbrake, by operating in a local context that is abstracted from the overall shape of either vehicle.
3.
Assignment of properties to groups of features. People are able to assign an abstract description to a whole set of shape features, and then make statements about the new abstraction, rather than simply about a single instance of it. This ability is particularly noticeable in its abscence - sufferers from visual agnosia are unable to construct abstract descriptions of sets of details in their visual field, and are thus unable to recognise generic classes of object, even though they can see all of the details.
4.
Qualitative size description and judgement. In many spatial reasoning situations, the absolute size of a given shape feature is not important. Its size relative to other shapes may be more important, as in the question ``will this washer fit over that bolt?'' Alternatively, its size may be altogether irrelevant, as in the question ``is that a bolt?''. If qualitative reasoning methods are available, it is possible to discuss relative size, or size-independent questions, without numerical information.



 
next up previous contents
Next: Representation of Detail Up: Two Methods for Qualitative Previous: Two Methods for Qualitative
Alan Blackwell
2000-11-17