Distributed coordination of autonomous agents by communicating on a need-to-know basis

Intelligent autonomous agents working cooperatively accomplish tasks more efficiently than single agents, as well as tasks that were infeasible for the single agent. For example, a single agent transporting a ton of blocks will need to take multiple trips, possibly even more trips than it can make, and will take longer than several agents transporting the blocks in parallel. Unexpected events can result in the need for agents to change their plans in order to adapt. During replanning, the expense of communication can hinder the amount of information passed. In this thesis, agents reduce the communication load while adapting to environmental events by examining what changes will be made to teammate plans and by passing information only on a need-to-know basis. More specifically, in this thesis, we describe a method whereby cooperating agents use identical planners and knowledge of the other agents' capabilities in order to pass information about the environmental conditions they observe that are necessary for their teammates to infer the correct actions to take. The agent must also pass conditions only to the teammates who are affected by the changes. Given that not all agents will have the same information about environmental conditions, the plans inferred by each agent may be different from their teammates. The Pertinent Information Planning (PIP) algorithm provided seeks to allow each agent to have an incorrect plan with respect to other agents but gives each agent a correct plan with respect to the actions that they perform. In order to determine the conditions that agents communicate, agents must repair their local plan and figure out which sets of actions have changed in the repaired plan. This thesis provides the

Department dc:contributor.department

Massachusetts Institute of Technology. Dept. of Electrical Engineering and Computer Science.

Grantor dc:publisher

Massachusetts Institute of Technology

Year dc:date.issued

2004

Author dc:creator

• Chen, Judy Y. (Judy Yann-Yun), 1980-

Advisor dc:contributor.advisor

• Brian C. Williams.

• Electrical Engineering and Computer Science.

Statement dc:rights

• M.I.T. theses are protected by copyright. They may be viewed from this source for any purpose, but reproduction or distribution in any format is prohibited without written permission. See provided URL for inquiries about permission.

Licence dc:rights.uri

• http://dspace.mit.edu/handle/1721.1/7582

Language dc:language.iso

eng

Handle dc:identifier.uri

http://hdl.handle.net/1721.1/17940

OAI identifier oai:identifier

oai:dspace.mit.edu:1721.1/17940