Approximate qualitative spatial reasoning

  • Authors:
  • Thomas Bittner;John G. Stell

  • Affiliations:
  • Department of Computer Science, Northwestern University, 1890 Maple Ave, Evanston, IL, 60201, U.S.A. (Fax: +1 (847) 491-5259, E-mail: bittner@cs.northwestern.edu);School of Computing University of Leeds, Leeds, LS2 9JT, U.K. (E-mail: jgs@comp.leeds.ac.uk)

  • Venue:
  • Spatial Cognition and Computation
  • Year:
  • 2001

Quantified Score

Hi-index 0.00

Visualization

Abstract

Qualitative relations between spatial regions play an importantrole in the representation and manipulation of spatial knowledge.The RCC5 and RCC8 systems of relations,used in the Region-Connection Calculus, are of fundamentalimportance. These two systems deal with ideal regions havingprecisely determined location. However,in many practical examples of spatial reasoning,regions are represented by finite approximations rather than known precisely.Approximations may be given by describing how a regionrelates to cells forming a partition of the space underconsideration. Although the RCC5 and RCC8 systems have beengeneralized to ``egg-yolk'' regions, in order to modelcertain types of vagueness, their extension to regionsapproximated in this way has not been discussed before.This paper presents two methods, the syntactic and the semantic, by which the RCC5 and RCC8 systemsmay be defined for approximate regions. The syntactic uses algebraicoperations on approximate regions which generalize operations on preciseregions. The semantic method makes use of the set of preciseregions which could be the intended interpretation of anapproximate region. Relationships between these two methods arediscussed in detail.alternative to navigation training with a map.