Bayesian Deconvolution of Signals Observed on Arrays

Ming Lin, Eric A. Suess, Robert H. Shumway, Rong Chen

Research output: Contribution to journalArticlepeer-review

Abstract

Time series data collected from arrays of seismometers are traditionally used to solve the core problems of detecting and estimating the waveform of a nuclear explosion or earthquake signal that propagates across the array. We consider here a parametric exponentially modulated autoregressive model. The signal is assumed to be convolved with random amplitudes following a Bernoulli normal mixture. It is shown to be potentially superior to the usual combination of narrow band filtering and beam forming. The approach is applied to analyzing series observed from an earthquake from Yunnan Province in China received by a seismic array in Kazakhstan.

Original languageEnglish (US)
Pages (from-to)837-850
Number of pages14
JournalJournal of Time Series Analysis
Volume37
Issue number6
DOIs
StatePublished - Nov 1 2016

All Science Journal Classification (ASJC) codes

  • Statistics and Probability
  • Statistics, Probability and Uncertainty
  • Applied Mathematics

Keywords

  • Nonlinear models, Bayes, Markov chain Monte Carlo, deconvolution, seismic arrays, nuclear monitoring

Fingerprint

Dive into the research topics of 'Bayesian Deconvolution of Signals Observed on Arrays'. Together they form a unique fingerprint.

Cite this