A Theory of Metric Spatial Inference

Drew McDermott

Efficient and robust spatial reasoning requires that the properties of real space be taken seriously. One approach to doing this is to assimilate facts into a "fuzzy map" of the positions and orientations of the objects involved in those facts. Then many inferences about distances and directions may be made by "just looking" at the map, to determine bounds on quantities of interest. For flexibility, there must be many frames of reference with respect to which coordinates are measure. The resulting representation supports many tasks, including finding routes from one place to another.


This page is copyrighted by AAAI. All rights reserved. Your use of this site constitutes acceptance of all of AAAI's terms and conditions and privacy policy.