Details on O-SBL(MCMC): A Compressive Sensing Algorithm for Sparse Signal Recovery for the SMV/MMV Problem Using Sparse Bayesian Learning and Markov Chain Monte Carlo Inference

Authors

Mohammad Shekaramiz, Utah State UniversityFollow
Todd K. Moon, Utah State UniversityFollow
Jacob H. Gunther, Utah State UniversityFollow

Document Type

Article

Publisher

Utah State University

Publication Date

2-11-2019

First Page

1

Last Page

10

Abstract

This report provides details on O-SBL(MCMC) algorithm for the recovery of jointly-sparse signals for the multiple measurement vector (MMV) problem. For the MMVs with this structure, the solution matrix, which is a collection of sparse vectors, is expected to exhibit joint sparsity across the columns. The notion of joint sparsity here means that the columns of the solution matrix share common supports. This algorithm employs a sparse Bayesian learning (SBL) model to encourage the joint sparsity structure across the columns of the solution. While the proposed algorithm is constructed for the MMV problems, it can also be applied to the single measurement vector (SMV) problems. Part of this work has been published in [1, 2].

Recommended Citation

Shekaramiz, Mohammad; Moon, Todd K.; and Gunther, Jacob H., "Details on O-SBL(MCMC): A Compressive Sensing Algorithm for Sparse Signal Recovery for the SMV/MMV Problem Using Sparse Bayesian Learning and Markov Chain Monte Carlo Inference" (2019). Electrical and Computer Engineering Faculty Publications. Paper 220.
https://digitalcommons.usu.edu/ece_facpub/220