Approximate spatial reasoning: Integrating qualitative and quantitative constraints

Abstract

Approximate reasoning refers in general to a broad class of solution techniques where either the inference procedure or the environment for inference is imprecise. Algorithms for approximate spatial reasoning are important for coping with the widespread imprecision and uncertainty in the real world. This paper develops an integrated framework for representing induced spatial constraints between a set of landmarks given imprecise, incomplete, and possibly conflicting quantitative and qualitative information about them. Fuzzy logic is used as the computational basis for both representing quantitative information and interpreting linguistically expressed qualitative constraints. © 1991.