Unlabelled Sensing: A Sparse Bayesian Learning Approach

TL;DR

This paper introduces novel Bayesian algorithms for recovering sparse signals and their permutation order in noisy, overcomplete measurement systems, incorporating correlated sparsity models and demonstrating improved performance through simulations.

Contribution

It presents new Bayesian update methods for jointly recovering sparse signals and their permutation order, including algorithms for correlated sparsity and permutation estimation.

Findings

Effective permutation recovery demonstrated via simulations.

Improved sparse signal reconstruction accuracy.

Algorithms handle correlated and uncorrelated sparsity models.

Abstract

We address the recovery of sparse vectors in an overcomplete, linear and noisy multiple measurement framework, where the measurement matrix is known upto a permutation of its rows. We derive sparse Bayesian learning (SBL) based updates for joint recovery of the unknown sparse vector and the sensing order, represented using a permutation matrix. We model the sparse matrix using multiple uncorrelated and correlated vectors, and in particular, we use the first order AR model for the correlated sparse vectors. We propose the Permutation-MSBL and a Kalman filtering based Permutation-KSBL algorithm for low-complex joint recovery of the uncorrelated and correlated sparse vectors, jointly with the permutation matrix. The novelty of this work emerges in providing a simple update step for the permutation matrix using the rearrangement inequality. We demonstrate the mean square error and the permutation recovery performance of the proposed algorithms using Monte Carlo simulations.