Constructing Diagrams to Support Situation Understanding and Planning: Part 1: Diagramming Group Motions
Abstract
Visual representations consisting of diagrammatic elements are ubiquitous in human problem solving. Diagrammatic Reasoning is a relatively new and challenging area of research in Artificial Intelligence and Human-Computer Interaction. The research described in this report is part of a larger project whose goal is to investigate a general Diagrammatic Reasoning architecture for problem solving. In some diagrammatic reasoning situations, such as military planning and weather prediction, it is necessary to abstract a mass of details into diagrammatic abstractions that are meaningful with respect to the problem solving goal. The research reported here focuses on this problem as it arises in a military domain. Commanders represent and monitor their situation understanding and plans by drawing lines, arrows, regions and other diagrammatic objects on maps that contain terrain and other mission-relevant information. Some of the diagrammatic objects are lines of motion, while other objects are regions that abstract information about occupancy, control, and so on, while yet others are point objects that abstract only the location of some entity. This report describes the issues involved in building a diagram extraction system for this domain. We describe an architecture for the generation of diagrams that abstract significant groups and represent their motions from information about the locations and movements of a large number of Blue and Red military entities engaged in action. We present experimental results applied to data from military exercises. We also discuss techniques needed to generate other types of diagrammatic objects, and outline our research objectives for the near future.
Document Details
- Document Type
- Technical Report
- Publication Date
- May 01, 2003
- Accession Number
- AD1001155
Entities
People
- Balasubramanian Chandrasekaran
- Bonny Banerjee
- John R. Josephson
- Robert Winkler
Organizations
- Ohio State University