(YIP) FOUNDATIONS OF ENSEMBLE ESTIMATION THEORY
Abstract
We propose in the project to establish foundations of ensemble estimation theory. Ensemble control deals with the problem of using a common control input to simultaneously steer a large population of individual systems. These systems are structurally identical, but show variations in system parameters. Because ensemble control requires minimal sensing and communication capabilities of the individual systems (or agents) and addresses the extreme scenario where there can be infinitely many agents, it naturally finds applications across various disciplines in engineering and science including unmanned aerial systems (e.g., swarms of drones), micro-robotic systems, nuclear magnetic resonance (NMR) spectroscopy, and laser and optical physics. However, many existing works assume implicitly that the initial states of the individual systems are known a priori to the controller and focus only on the control part. The literature is very sparse on the counterpart of ensemble control, namely ensemble estimation. In a nutshell, ensemble estimation is about using a single, finite-dimensional measurement output to estimate the initial state of every individual system in the ensemble. This is a completely new research area. In fact, observability of a nonlinear ensemble system has not been addressed until recently by the PI. Potential benefits of ensemble systems are limited by our lack of knowledge in ensemble estimation theory. The gap needs to be filled imminently.
Document Details
- Document Type
- DoD Grant Award
- Publication Date
- Aug 12, 2021
- Source ID
- FA95502010076
Entities
People
- Xudong Chen
Organizations
- Air Force Office of Scientific Research
- Regents of the University of Colorado
- United States Air Force