Structure-Aware Bayesian Compressive Sensing for Frequency-Hopping Spectrum Estimation with Missing Observations

TL;DR

This paper introduces a structure-aware Bayesian compressive sensing approach that leverages the inherent frequency-hopping signal structure within a time-frequency framework to accurately estimate spectra despite missing data, outperforming existing methods.

Contribution

It proposes a novel structure-aware Bayesian compressive sensing algorithm combined with a tailored time-frequency kernel for robust spectrum estimation of frequency-hopping signals with missing observations.

Findings

High-resolution spectrum estimation with significant data missing

Effective suppression of cross-terms and artifacts

Superiority over existing spectrum estimation techniques

Abstract

In this paper, we address the problem of spectrum estimation of multiple frequency-hopping (FH) signals in the presence of random missing observations. The signals are analyzed within the bilinear time-frequency (TF) representation framework, where a TF kernel is designed by exploiting the inherent FH signal structures. The designed kernel permits effective suppression of cross-terms and artifacts due to missing observations while preserving the FH signal auto-terms. The kernelled results are represented in the instantaneous autocorrelation function domain, which are then processed using a re-designed structure-aware Bayesian compressive sensing algorithm to accurately estimate the FH signal TF spectrum. The proposed method achieves high-resolution FH signal spectrum estimation even when a large portion of data observations is missing. Simulation results verify the effectiveness of the proposed method and its superiority over existing techniques.