Distributed Coordination in Multi-Agent Networked Systems: Algorithms and Fundamental Limits
Abstract
Multi-agent networked systems arise in many application domains that are of extremeimportance to the Air Force, such as reconnaissance and surveillance, autonomous operation ofUAVs, sensor coverage, distributed cooperative estimation, vehicle targeting, among manyothers. Regardless of the specific application domain, the central goal is to achieve desiredcollective behaviors of the networked system through the design of admissible local agentcontrol algorithms. This is nontrivial especially due to challenges related to local connectivity,imperfect communication, and time-varying, uncertain environments. Consequently, research hasevolved in a per-application manner. Despite the large number of algorithms that have beendeveloped for various applications, we lack the knowledge and methods required to deliberatelyand systematically shape the network collective behavior. Consequently, the existing theory inmulti-agent coordination is not well prepared for dealing with the growing complexity associatedwith Air Force systems.
Document Details
- Document Type
- DoD Grant Award
- Publication Date
- Apr 09, 2018
- Source ID
- FA95501810150
Entities
People
- Na Li
Organizations
- Air Force Office of Scientific Research
- President and Fellows of Harvard College
- United States Air Force