Compressive Sensing Matrix Design and Evaluation with Cognitive Radio Applications
Abstract
This dissertation is aimed at evaluating a sensing matrix in an overcomplete compressive sensing (CS) system. The main theoretical problems addressed are: Given a redundant (overcomplete)dictionary, a sensing matrix, and a greedy approximation method such as orthogonal matching pursuit, what tools can ascertain the suitability of a given sensing matrix, in the course of performing CS with the dictionary? In general, how does the relationship between the sensing matrix and dictionary (regarding structure/composition) affect the greedy support estimation of a sparse signal?
Document Details
- Document Type
- Technical Report
- Publication Date
- Oct 25, 2018
- Accession Number
- AD1086330
Entities
People
- John J. Kelly
Organizations
- Rome Laboratory