Decentralized Estimation with Dependent Gaussian Observations

Fangrong Peng, Biao Chen

Research output: Contribution to journalArticle

1 Citation (Scopus)

Abstract

This paper considers decentralized estimation with correlated noises under the Bayesian framework. For a tandem network with correlated additive Gaussian noises, we establish that threshold quantizers on local observations are optimal in the sense of maximizing Fisher information at the fusion center; this is true despite the fact that subsequent estimators may differ at the fusion center, depending on the statistical distribution of the parameter to be estimated. In addition, it is always beneficial to have the better sensor, i.e., the one with higher signal-To-noise ratio, serve as the fusion center in a tandem network. Finally, we identify different correlation regimes in terms of their impact on the estimation performance. These include the well-known case where negatively correlated noise benefits estimation performance as it facilitates noise cancellation, as well as two distinct regimes with positively correlated noises.

Original languageEnglish (US)
Article number7752980
Pages (from-to)1172-1182
Number of pages11
JournalIEEE Transactions on Signal Processing
Volume65
Issue number5
DOIs
StatePublished - Mar 1 2017

Fingerprint

Fusion reactions
Signal to noise ratio
Sensors

Keywords

  • communication direction
  • correlated observations
  • Decentralized estimation
  • quantizer design
  • tandem network

ASJC Scopus subject areas

  • Signal Processing
  • Electrical and Electronic Engineering

Cite this

Decentralized Estimation with Dependent Gaussian Observations. / Peng, Fangrong; Chen, Biao.

In: IEEE Transactions on Signal Processing, Vol. 65, No. 5, 7752980, 01.03.2017, p. 1172-1182.

Research output: Contribution to journalArticle

@article{178a8d074bd94670b5f10b3972595a47,
title = "Decentralized Estimation with Dependent Gaussian Observations",
abstract = "This paper considers decentralized estimation with correlated noises under the Bayesian framework. For a tandem network with correlated additive Gaussian noises, we establish that threshold quantizers on local observations are optimal in the sense of maximizing Fisher information at the fusion center; this is true despite the fact that subsequent estimators may differ at the fusion center, depending on the statistical distribution of the parameter to be estimated. In addition, it is always beneficial to have the better sensor, i.e., the one with higher signal-To-noise ratio, serve as the fusion center in a tandem network. Finally, we identify different correlation regimes in terms of their impact on the estimation performance. These include the well-known case where negatively correlated noise benefits estimation performance as it facilitates noise cancellation, as well as two distinct regimes with positively correlated noises.",
keywords = "communication direction, correlated observations, Decentralized estimation, quantizer design, tandem network",
author = "Fangrong Peng and Biao Chen",
year = "2017",
month = "3",
day = "1",
doi = "10.1109/TSP.2016.2631463",
language = "English (US)",
volume = "65",
pages = "1172--1182",
journal = "IEEE Transactions on Signal Processing",
issn = "1053-587X",
publisher = "Institute of Electrical and Electronics Engineers Inc.",
number = "5",

}

TY - JOUR

T1 - Decentralized Estimation with Dependent Gaussian Observations

AU - Peng, Fangrong

AU - Chen, Biao

PY - 2017/3/1

Y1 - 2017/3/1

N2 - This paper considers decentralized estimation with correlated noises under the Bayesian framework. For a tandem network with correlated additive Gaussian noises, we establish that threshold quantizers on local observations are optimal in the sense of maximizing Fisher information at the fusion center; this is true despite the fact that subsequent estimators may differ at the fusion center, depending on the statistical distribution of the parameter to be estimated. In addition, it is always beneficial to have the better sensor, i.e., the one with higher signal-To-noise ratio, serve as the fusion center in a tandem network. Finally, we identify different correlation regimes in terms of their impact on the estimation performance. These include the well-known case where negatively correlated noise benefits estimation performance as it facilitates noise cancellation, as well as two distinct regimes with positively correlated noises.

AB - This paper considers decentralized estimation with correlated noises under the Bayesian framework. For a tandem network with correlated additive Gaussian noises, we establish that threshold quantizers on local observations are optimal in the sense of maximizing Fisher information at the fusion center; this is true despite the fact that subsequent estimators may differ at the fusion center, depending on the statistical distribution of the parameter to be estimated. In addition, it is always beneficial to have the better sensor, i.e., the one with higher signal-To-noise ratio, serve as the fusion center in a tandem network. Finally, we identify different correlation regimes in terms of their impact on the estimation performance. These include the well-known case where negatively correlated noise benefits estimation performance as it facilitates noise cancellation, as well as two distinct regimes with positively correlated noises.

KW - communication direction

KW - correlated observations

KW - Decentralized estimation

KW - quantizer design

KW - tandem network

UR - http://www.scopus.com/inward/record.url?scp=85012933769&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85012933769&partnerID=8YFLogxK

U2 - 10.1109/TSP.2016.2631463

DO - 10.1109/TSP.2016.2631463

M3 - Article

AN - SCOPUS:85012933769

VL - 65

SP - 1172

EP - 1182

JO - IEEE Transactions on Signal Processing

JF - IEEE Transactions on Signal Processing

SN - 1053-587X

IS - 5

M1 - 7752980

ER -