Integrated Sensing and Processing (ISP). A Mathematical Methodology for Managing and Integrating Sensors and Processors in Distributed Systems for Radar and Communication
Abstract
The objective of this effort is to develop tools for integrating sensing and processing over as wide a range of application areas as possible. The approach is to consider systems of targets and sensors in as general a general mathematical formulation as possible, to develop mathematical tools to study such systems, and to apply the tools to problems in radar and communications. Accomplishments include results on the characterization of sources and sensors and decision-directed sensing and processing implemented through partially observed Markov decision processes (POMDPs) and binary hypertrees (BHCs). The characterization of sensors and sources shows that time-frequency distributions and wireless scattering functions can both be estimated by as a convolution of a Rihaczek time-frequency density with a time-frequency kernel function. POMDPs have been applied to a sensor-scheduling algorithm, and results indicate that the use of POMDPs may lead to much more efficient sensor network management. BHCs have been applied for the efficient solution of classifier problems.
Document Details
- Document Type
- Technical Report
- Publication Date
- Apr 01, 2005
- Accession Number
- ADA441567
Entities
People
- Chad M. Spooner