Knowledge-aided hyperparameter-free Bayesian detection in stochastic homogeneous environments

Pu Wang, Hongbin Li, Olivier Besson, Jun Fang

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

1 Scopus citations

Abstract

This paper considers adaptive signal detection in stochastic homogeneous environments where the disturbance covariance matrix of both test and training signals, R, is assumed to be a random matrix with a priori knowledge of R. Unlike existing detectors assuming a known hyperparameter associated with R, a knowledge-aided detector with the capability of automatic weighting is considered by accounting for the uncertainty of the prior knowledge. Specifically, the generalized likelihood ratio test (GLRT) is utilized to develop the test statistic, along with the maximum marginal likelihood (MML) estimation of the hyperparameter. The proposed KA-MML-GLRT detector is evaluated by numerical simulations and the results show improved detection performance over conventional and knowledge-aided detectors, especially in the case of limited training signals and inaccurate prior knowledge.

Original languageEnglish
Title of host publication2016 IEEE International Conference on Acoustics, Speech and Signal Processing, ICASSP 2016 - Proceedings
Pages2901-2905
Number of pages5
ISBN (Electronic)9781479999880
DOIs
StatePublished - 18 May 2016
Event41st IEEE International Conference on Acoustics, Speech and Signal Processing, ICASSP 2016 - Shanghai, China
Duration: 20 Mar 201625 Mar 2016

Publication series

NameICASSP, IEEE International Conference on Acoustics, Speech and Signal Processing - Proceedings
Volume2016-May
ISSN (Print)1520-6149

Conference

Conference41st IEEE International Conference on Acoustics, Speech and Signal Processing, ICASSP 2016
Country/TerritoryChina
CityShanghai
Period20/03/1625/03/16

Keywords

  • Stochastic homogeneous model
  • generalized likelihood ratio test
  • maximum marginal likelihood estimation

Fingerprint

Dive into the research topics of 'Knowledge-aided hyperparameter-free Bayesian detection in stochastic homogeneous environments'. Together they form a unique fingerprint.

Cite this