Spectral Super-Resolution With Prior Knowledge

Kumar Vijay Mishra, Myung Cho, Anton Kruger, Weiyu Xu

Research output: Contribution to journalArticle

24 Citations (Scopus)

Abstract

We address the problem of super-resolution frequency recovery using prior knowledge of the structure of a spectrally sparse, undersampled signal. In many applications of interest, some structure information about the signal spectrum is often known. The prior information might be simply knowing precisely some signal frequencies or the likelihood of a particular frequency component in the signal. We devise a general semidefinite program to recover these frequencies using theories of positive trigonometric polynomials. Our theoretical analysis shows that, given sufficient prior information, perfect signal reconstruction is possible using signal samples no more than thrice the number of signal frequencies. Numerical experiments demonstrate great performance enhancements using our method. We show that the nominal resolution necessary for the grid-free results can be improved if prior information is suitably employed.

Original languageEnglish (US)
Article number7145484
Pages (from-to)5342-5357
Number of pages16
JournalIEEE Transactions on Signal Processing
Volume63
Issue number20
DOIs
StatePublished - Oct 15 2015

Fingerprint

Signal reconstruction
Polynomials
Recovery
Experiments

All Science Journal Classification (ASJC) codes

  • Signal Processing
  • Electrical and Electronic Engineering

Cite this

Mishra, Kumar Vijay ; Cho, Myung ; Kruger, Anton ; Xu, Weiyu. / Spectral Super-Resolution With Prior Knowledge. In: IEEE Transactions on Signal Processing. 2015 ; Vol. 63, No. 20. pp. 5342-5357.
@article{8b1e84ed4c834277b04070a439e17dd7,
title = "Spectral Super-Resolution With Prior Knowledge",
abstract = "We address the problem of super-resolution frequency recovery using prior knowledge of the structure of a spectrally sparse, undersampled signal. In many applications of interest, some structure information about the signal spectrum is often known. The prior information might be simply knowing precisely some signal frequencies or the likelihood of a particular frequency component in the signal. We devise a general semidefinite program to recover these frequencies using theories of positive trigonometric polynomials. Our theoretical analysis shows that, given sufficient prior information, perfect signal reconstruction is possible using signal samples no more than thrice the number of signal frequencies. Numerical experiments demonstrate great performance enhancements using our method. We show that the nominal resolution necessary for the grid-free results can be improved if prior information is suitably employed.",
author = "Mishra, {Kumar Vijay} and Myung Cho and Anton Kruger and Weiyu Xu",
year = "2015",
month = "10",
day = "15",
doi = "10.1109/TSP.2015.2452223",
language = "English (US)",
volume = "63",
pages = "5342--5357",
journal = "IEEE Transactions on Signal Processing",
issn = "1053-587X",
publisher = "Institute of Electrical and Electronics Engineers Inc.",
number = "20",

}

Spectral Super-Resolution With Prior Knowledge. / Mishra, Kumar Vijay; Cho, Myung; Kruger, Anton; Xu, Weiyu.

In: IEEE Transactions on Signal Processing, Vol. 63, No. 20, 7145484, 15.10.2015, p. 5342-5357.

Research output: Contribution to journalArticle

TY - JOUR

T1 - Spectral Super-Resolution With Prior Knowledge

AU - Mishra, Kumar Vijay

AU - Cho, Myung

AU - Kruger, Anton

AU - Xu, Weiyu

PY - 2015/10/15

Y1 - 2015/10/15

N2 - We address the problem of super-resolution frequency recovery using prior knowledge of the structure of a spectrally sparse, undersampled signal. In many applications of interest, some structure information about the signal spectrum is often known. The prior information might be simply knowing precisely some signal frequencies or the likelihood of a particular frequency component in the signal. We devise a general semidefinite program to recover these frequencies using theories of positive trigonometric polynomials. Our theoretical analysis shows that, given sufficient prior information, perfect signal reconstruction is possible using signal samples no more than thrice the number of signal frequencies. Numerical experiments demonstrate great performance enhancements using our method. We show that the nominal resolution necessary for the grid-free results can be improved if prior information is suitably employed.

AB - We address the problem of super-resolution frequency recovery using prior knowledge of the structure of a spectrally sparse, undersampled signal. In many applications of interest, some structure information about the signal spectrum is often known. The prior information might be simply knowing precisely some signal frequencies or the likelihood of a particular frequency component in the signal. We devise a general semidefinite program to recover these frequencies using theories of positive trigonometric polynomials. Our theoretical analysis shows that, given sufficient prior information, perfect signal reconstruction is possible using signal samples no more than thrice the number of signal frequencies. Numerical experiments demonstrate great performance enhancements using our method. We show that the nominal resolution necessary for the grid-free results can be improved if prior information is suitably employed.

UR - http://www.scopus.com/inward/record.url?scp=84941927058&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84941927058&partnerID=8YFLogxK

U2 - 10.1109/TSP.2015.2452223

DO - 10.1109/TSP.2015.2452223

M3 - Article

AN - SCOPUS:84941927058

VL - 63

SP - 5342

EP - 5357

JO - IEEE Transactions on Signal Processing

JF - IEEE Transactions on Signal Processing

SN - 1053-587X

IS - 20

M1 - 7145484

ER -