Estimation, Control, and Redundancy Management for Uncertain Networks of Cooperating Agents
Abstract
The coordination of spatially distributed systems of cooperating agents, which perform an assigned mission in the presence of uncertainty and system faults, is an important emerging technology. The actions and health of these distributed systems depend upon the information that can be communicated and the knowledge of the current capabilities of all cooperating agents. Methodologies for the distribution of estimation and redundancy management functions over the dynamic network of cooperating agents were developed, leading to effective team strategies Progress has been made on various aspects of the distributed systems problem. From the fundamental level we investigated the decentralized control problem with constrained communication. In parallel the allocation of transmit power in wireless networks was a focus of study into the decentralized control problem because it has a simple structure and the information communicated is constrained. In the area of health monitoring new robust analytical redundancy methods have been developed which detects, identifies, and reconstructs sensor, actuator and plant faults. A robust multiple-fault filter is developed based on a performance measure from which the desired detection subspaces are approximately constructed. This detection filter formulation, which includes uncertainty, is the bases for single-fault time-varying, decentralized detection filters, and fault magnitude reconstruction. An innovative application of distributed detection filters methodology is to the target track association problem. Finally, the distributed estimation problem was addressed by considering elements of the relative navigation problem among distributed vehicles.
Document Details
- Document Type
- Technical Report
- Publication Date
- Jan 31, 2003
- Accession Number
- ADA416352
Entities
People
- Jason Speyer
Organizations
- University of California, Los Angeles