A Fast Bayesian Model for Latent Radio Signal

Brett Houlding; Arnab Bhattacharya; Simon Wilson; Tim Forde

5th International Workshop on Spatial Stochastic Models for Wireless Networks

Research Article

A Fast Bayesian Model for Latent Radio Signal

Cite: BibTeX Plain Text

@INPROCEEDINGS{10.1109/WIOPT.2009.5291568,
    author={Brett Houlding and Arnab Bhattacharya and Simon Wilson and Tim Forde},
    title={A Fast Bayesian Model for Latent Radio Signal},
    proceedings={5th International Workshop on Spatial Stochastic Models for Wireless Networks},
    publisher={IEEE},
    proceedings_a={SPASWIN},
    year={2009},
    month={10},
    keywords={Dynamic spectrum access latent radio signal Bayesian estimation integrated nested Laplace approximation},
    doi={10.1109/WIOPT.2009.5291568}
}

Brett Houlding
Arnab Bhattacharya
Simon Wilson
Tim Forde
Year: 2009
A Fast Bayesian Model for Latent Radio Signal
SPASWIN
IEEE
DOI: 10.1109/WIOPT.2009.5291568

Brett Houlding¹^,*, Arnab Bhattacharya¹^,*, Simon Wilson¹^,*, Tim Forde¹^,*

1: CTVR, Department of Statistics, Trinity College Dublin, Ireland

*Contact email: houldinb@tcd.ie, bhattaca@tcd.ie, swilson@tcd.ie, timforde@mee.tcd.ie

Abstract

This paper considers the use of a recently developed Bayesian statistical approximation technique that leads to very fast determination of highly accurate estimates for latent radio signal power. Following suitable data analysis, a first order non-stationary auto-regressive process is considered for latent radio signal and the fast approximation technique is then used to provide accurate estimates of the hidden model parameters. These estimates are based on having received several noisy, but spatially correlated, observations of the true latent signal. The implication of this technique for real time decision analysis and the problem of finding, and making use of, so-called radio spectrum holes is also discussed.

Keywords: Dynamic spectrum access latent radio signal Bayesian estimation integrated nested Laplace approximation

Published: 2009-10-23
Publisher: IEEE
Modified: 2010-05-16

: http://dx.doi.org/10.1109/WIOPT.2009.5291568