Coordinating very large groups of wide area search munitions
Abstract
Coordinating hundreds or thousands of unmanned aerial vehicles (UAVs), presents a variety of new exciting challenges, over and above the challenges of building single UAVs and small teams of UAVs. We are specifically interested in coordinating large groups of Wide Area Search Munitions (WASMs), which are part UAV and part munition. We are developing a "flat", distributed organization to provide the robustness and flexibility required by a group where team members will frequently leave. Building on established teamwork theory and infrastructure we are able to build large teams that can achieve complex goals using completely distributed intelligence. However, as the size of the team is increased, new issues arise that require novel algorithms. Specifically, key algorithms that work well for relatively small teams, fail to scale up to very large teams. We have developed novel algorithms meeting the requirements of large teams for the tasks of instantiating plans, sharing information and allocating roles. We have implemented these algorithms in reusable software proxies using the novel design abstraction of a coordination agent that encapsulates a piece of coordination protocol. We illustrate the effectiveness of the approach with 200 WASMs coordinating to find and destroy ground based targets in support of a manned aircraft.
forthcoming
BibTeX
@article{Scerri-2004-16963,author = {Paul Scerri and Elizabeth Liao and Y. Xu and and G. Lai and Katia Sycara},
title = {Coordinating very large groups of wide area search munitions},
journal = {Theory and Algorithms for Cooperative Systems},
year = {2004},
month = {August},
pages = {451 - 480},
}