Agent-Based Support for Human/Agent Teams - Robotics Institute Carnegie Mellon University

Agent-Based Support for Human/Agent Teams

Terence Payne, Terri L. Lenox, Susan K. Hahn, and Katia Sycara
Conference Paper, Proceedings of CHI '00 Extended Abstracts on Human Factors in Computing Systems (CHI EA '00), pp. 22 - 23, April, 2000

Abstract

In this paper, we present an interface agent, MokSAF, which facilitates time-critical team-planning tasks for teams of both humans and heterogeneous software agents. This agent assists in the formation of teams of humans (via other MokSAF agents) and task agents that can autonomously perform team subtasks. It provides a suitable interaction mechanism to instruct the various task agents in the team; and, by monitoring the human's progress, reallocate or modify the sub-tasks if the human fails to achieve that sub-task. A military domain has been used to investigate this interface agent. The task consists of three military (human) commanders that each assemble a platoon, and plan routes so that all three platoons arrive at a given rendezvous by a specified time. An experimental study has been conducted to evaluate MokSAF and the assistance provided by one of three different task agents, and the results summarized.

BibTeX

@conference{Payne-2000-8000,
author = {Terence Payne and Terri L. Lenox and Susan K. Hahn and Katia Sycara},
title = {Agent-Based Support for Human/Agent Teams},
booktitle = {Proceedings of CHI '00 Extended Abstracts on Human Factors in Computing Systems (CHI EA '00)},
year = {2000},
month = {April},
pages = {22 - 23},
publisher = {ACM Press},
address = {New York, NY},
}