Doctoral Dissertations

Author

Nathan Eloe

Keywords and Phrases

Qualitative Spatial Reasoning; Region Connection Calculi (RCC); Temporal Reasoning

Abstract

"Qualitative Spatial Reasoning (QSR) has varying applications in Geographic Information Systems (GIS), visual programming language semantics, and digital image analysis. Systems for spatial reasoning over a set of objects have evolved in both expressive power and complexity, but implementations or usages of these systems are not common. This is partially due to the computational complexity of the operations required by the reasoner to make informed decisions about its surroundings. These theoretical systems are designed to focus on certain criteria, including efficiency of computation, ease of human comprehension, and expressive power. Sadly, the implementation of these systems is frequently left as an exercise for the reader.

Herein, a new QSR system, VRCC-3D+, is proposed that strives to maximize expressive power while minimizing the complexity of reasoning and computational cost of using the system. This system is an evolution of RCC-3D; the system and implementation are constantly being refined to handle the complexities of the reasoning being performed. The refinements contribute to the accuracy, correctness, and speed of the implementation. To improve the accuracy and correctness of the implementation, a way to dynamically change error tolerance in the system to more accurately reflect what the user sees is designed. A method that improves the speed of determining spatial relationships between objects by using composition tables and decision trees is introduced, and improvements to the system itself are recommended; by streamlining the relation set and enforcing strict rules for the precision of the predicates that determine the relationships between objects. A potential use case and prototype implementation is introduced to further motivate the need for implementations of QSR systems, and show that their use is not precluded by computational complexity."--Abstract, page iv.

Advisor(s)

Leopold, Jennifer
Sabharwal, Chaman

Committee Member(s)

Jiang, Wei
Yin, Zhaozheng
Insall, Matt

Department(s)

Computer Science

Degree Name

Ph. D. in Computer Science

Publisher

Missouri University of Science and Technology

Publication Date

Spring 2015

Journal article titles appearing in thesis/dissertation

  • Efficient computation of object boundary intersection and error tolerance in VRCC-3D+
  • Efficient determination of spatial relations using composition tables and decision trees
  • Spatial temporal reasoning using image processing, physics, and qualitative spatial reasoning
  • Dual graph partitioning for bottom-up BVH construction
  • A more efficient representation of obscuration for VRCC-3D+ relations

Pagination

xv, 117 pages

Note about bibliography

Includes bibliographic references.

Rights

© 2015 Nathan Webster Eloe, All rights reserved.

Document Type

Dissertation - Open Access

File Type

text

Language

English

Subject Headings

Qualitative reasoningSpatial analysis (Statistics)Space and time -- Mathematical modelsLogic, Symbolic and mathematicalKnowledge representation (Information theory)

Thesis Number

T 10711

Electronic OCLC #

913390384

Share

 
COinS