[BibTeX] [RIS]
Using orientation information for qualitative spatial reasoning
Type of publication: Incollection
Citation: freksa_92_using
Publication status: Published
Booktitle: Theories and methods of spatio-temporal reasoning in geographic space
Series: Lecture Notes in Computer Science
Volume: 639
Year: 1992
Pages: 162-178
Publisher: Springer, Berlin
DOI: 10.1007/3-540-55966-3_10
Abstract: A new approach to representing qualitative spatial knowledge and to spatial reasoning is presented: This approach is motivated by cognitive considerations and is based on relative orientation information about spatial environments. The approach aims at exploiting properties of physical space which surface when the spatial knowledge is structured according to conceptual neighborhood of spatial relations.The paper introduces the notion of conceptual neighborhood and its relevance for qualitative temporal reasoning. The extension of the benefits to spatial reasoning is suggested. Several approaches to qualitative spatil reasoning are briefly reviewed. Differences between the temporal and the spatial domain are outlined. A way of transferring a qualitative temporal reasoning oriented representation and reasoning approach is presented and illustrated. An example for an application of the approach is discussed.
Userfields: content={Abstract 1. Introduction 1.1 Background 1.2 Qualitative Reasoning 1.3 Spatial Reasoning 1.4 Existing Approaches to Qualitative Spatial Reasoning 2. Qualitative Orientation 2.1 Dimensionality of Space and Domain-Inherent Constraints 2.2 Orientation Values and Properties of Qualitative Oriemtation 2.3 Augmenting Qualitative Orientation Relations 3. Conceptual Neighborhood and Spatiel Knowledge 3.1 Conceptual Neighborhood of Spatial Relations 3.2 What are Appropriate Entities to be Spatially Related? 4. Qualitative Spatial Reasoning 4.1 Orientation-Based Inferences 4.2 Higher Resolution Reasoning 4.3 Applications 5. Discussion}, date-added={2012-09-03 15:47:30 +0200}, date-modified={2012-09-03 15:47:30 +0200}, project={fremdliteratur}, registry={A67}, state={copied; read; summarized}, summary={1. Introduction In this chapter a general motivation for qualitative reasoning and the presented approach is given. Some general issues about spatial reasoning are discussed and the existing literature on qualitative spatial reasoning is shortly reviewed. 2. Qualitative Orientation Here the use of relative orientation information is motivated and the fundamental principles behind the idea of a \"relation space\" are shown. Then directional orientation ( denoted by an oriented line or an odered set of two points ) and relative orientation ( given by two oriented lines or two ordered set of points) are defined. Orientation relations (same, opposite,left,right) are derived from this sceme and their properties discussed. The introduction of a front/back discrimination is motivated which results in 8 disjoint orientation relations. 4 different ways in which the front/back dichomoty can be apllied to two vectors ab and bc are suggested. Merging of the two possibilities for one vector results in a qualitative description of any location c wrt. location b and orientation ab and wrt. location a and orietation ba (15 qualitative locations). 3. Conceptual Neighborhood and Spatial Knowledge Conceptual neighborhood for the relations of chapter 2 is introduced and and its advantages examined. Then the use of points as appropriate entities to be spatially related is motivated. 4. Qualitative Spatial Reasoning This chapter shows how the qualitative spatial relation of a vector bd to a vector ab can be infered when the relations from bc to ab and from cd to bc are known using the simple case of a single front/back dichomoty (see chapter 2). A general composition table for this kind of inference is presented (which is arranged in a way that conceptually neighboring relations are at neighboring locations in the tabe). A simple formula for computing the infered relations (numbered from 0 to 7) is given. It is briefly mentioned how the results can be used for further inferences. Then a composition table for the more complex case of two front/back dichomoties (see chapter 2) is presented resulting in more precise inferences. It is also suggested how inferences can be refined by evidence from multiple sources. Finally possible applications of such a kind of orientation-based qualitative spatial reasoning are listed: localization in space, wayfinding and route comparation. 5. Discussion It is argued that the presented approach is cognitively adequate and some advantages of neighborhood-based reasoning are given. Finally it is shown that other kinds of inferences are also possible, for example: infer d related to ac, b related to ac or b related to ad instead of d related to ab in chapter 4. For such inferences new composition tables are required which share important properties with the one discussed in this article.},
Authors Freksa, Christian
Editors Frank, Andrew U.
Campari, Irene
Formentini, Ubaldo
  • Pisa 20h.pdf