Modeling Directional Knowledge and Reasoning in Environmental Space: Testing Qualitative Metrics

Abstract

Researchers from a variety of disciplines have proposed models of human spatial knowledge and reasoning in order to explain spatial behavior in environmental spaces, such as buildings, neighborhoods, and cities. A common component of these models is a set of hypotheses about the geometry of spatial knowledge, particularly with respect to the roles of topological and metric knowledge. Recently, mathematicians and computer scientists interested in formally modeling everyday intelligent spatial behavior have developed models incorporating “qualitative” spatial reasoning (“naive” spatial reasoning). One branch of this effort has been the development of so-called “qualitative metric” models to solve problems such as wayfinding. A qualitative metric employs more sophisticated geometry than just topology but at a relatively imprecise or coarse-grained level. Such models essentially reason with a small finite number of quantitative categories for direction and/or distance. In this chapter, we evaluate the abilities of qualitative metric models to account for human knowledge of directions by comparing simulations derived from qualitative metrics to empirical data and theorizing derived from human-subjects testing.