Distributed Learning and Information Dynamics In Networked Autonomous Systems
Abstract
This project is motivated by the need to enable advanced operations of teams of autonomous vehicles to learn and adapt to uncertain and hostile environments under effective utilization of communications resources. Of particular interest is the interplay between distributed learning and information dynamics. Distributed learning refers to a collection of interacting agents with limited local processing, information, and communications, all seeking to achieve a global objective in an uncertain and possibly hostile environment. Information dynamics refers to the architecture, either inherited or designed, of information flow among the distributed agents. The interplay of distributed learning algorithms and information dynamics can have dramatic effects on the efficiency of the collective. Specific research thrusts include: i) online resource allocation, ii) networked operations, and iii) evolving and uncertain operations.
Document Details
- Document Type
- Technical Report
- Publication Date
- Nov 20, 2015
- Accession Number
- AD1000659
Entities
People
- Asuman Özdağlar
- Emilio Frazzoli
- Eyad H. Abed
- Jeff Shamma
- John Baras
- Leslie Kaebling
- Nina Balcan
- Nuno C. Martins
- Peyton Young
- Éric Féron
Organizations
- Georgia Tech Research Corporation