In recent years qualitative reasoning approaches have become increasingly popular and are preferred over quantitative numeric approaches for applications in the field of AI. This is due to several factors. The most striking argument in the field of temporal and spatial reasoning is that humans are not able to give precise numeric estimates of their environment, e.g., if asked to estimate temporal duration or object size. Nevertheless we are capable of dealing with our surrounding world in a very efficient manner and are able to produce qualitative descriptions of it. In the field of point like measures, such as object dimensions or the duration of intervals, only a few new qualitative approaches have been developed, such as, Order of Magnitude for technical domains, and thus researchers tend to stick with numeric approaches. In this paper we present a new approach based on cognitive considerations of how humans perceive spatial dimensions and how they reason with this spatial knowledge. We then describe how reasoning is performed within the new calculus and how it can be adopted for representing not only one-dimensional measures, but also areas, volumes and proportions.
展开▼
