Final Report: Minimax Compressed Sensing Reconstruction
Abstract
In compressive sensing, one basic issue is the robustness of signal recovery solutions in the presence of uncertainties. The main objective of this project is to analysis the robustness of compressive sensing solutions, and derive, through minimax optimization, solutions that are robust to uncertainties (or perturbations) in modeling and in measurements. Exact solutions of compressive sensing solutions to perturbations were obtained. Algorithms for sensitivity reduction in sparse signal recovery solutions we designed. Algorithms for obtaining robust compressive sensing solutions under the worst-case perturbations were obtained through the Alternating Direction Method of Multipliers. Finally, the optimality of Wiener filter was established under non-Gaussian distributions of signals.
Document Details
- Document Type
- Technical Report
- Publication Date
- Sep 30, 2016
- Accession Number
- AD1063096
Entities
People
- Dror Baron
- Hamid Krim
- Liyi Dai
Organizations
- North Carolina State University