Accession Number:

ADA636691

Title:

Testbed for Intelligent Aiding Using the LOCATE Workspace Layout Tool

Descriptive Note:

Final rept.

Corporate Author:

ARTIFICIAL INTELLIGENCE MGT AND DEVELOPMENT CORP (AIMDC) TORONTO (ONTARIO)

Personal Author(s):

Report Date:

1999-11-01

Pagination or Media Count:

67.0

Abstract:

The purpose of the work reported here was to develop generalizable principles and techniques for intelligent aiding useful to projects developed at Canadas Defence and Civil Institute of Environmental Medicine DCIEM. In realizing that purpose, the principles and techniques of intelligent aiding implemented in DCIEMs LOCATE Workspace Layout Tool were combined with other, similar principles and techniques from related areas of research. LOCATE is a CAD tool that allows users to create designs and analyze their communication efficiency. The LOCATE tool is serving as the focus for discussion, design and implementation in an emerging testbed for elaborating ideas on intelligent aiding. Earlier work on building an infrastructure to support such aiding in LOCATE was extended so that the tracking of all low-level user actions is now complete, as is the ability of LOCATE to infer the goals those actions might imply. Further extensions included the identification and elaboration of categories of actions that imply a need for help. A category of Non-Responsive Actions, or actions for which there is no system response, was selected for implementation. Help was designed and implemented, first of all, to inform users that such actions produce no response and second, to identify and help users achieve whatever goals they may have had in mind when they performed those actions. Task and help, actions and goals are represented in explicit models of task, user and system in LOCATE, and are displayed in associated windows in its interface. In the Task Model, an action and goal history is maintained as the user creates and analyzes designs. As that work proceeds, help of various kinds is initiated by LOCATE and, in a similar way, its System Model maintains an action and goal help history. Several projects were identified as likely candidates for the application of the principles and techniques identified, but none was at a stage where such application was possible.

Subject Categories:

  • Computer Programming and Software

Distribution Statement:

APPROVED FOR PUBLIC RELEASE