Particle filtering with path sampling and an application to a bimodal ocean current model

Jonathan Weare

Research output: Contribution to journalArticle

Abstract

This paper introduces a recursive particle filtering algorithm designed to filter high dimensional systems with complicated non-linear and non-Gaussian effects. The method incorporates a parallel marginalization (PMMC) step in conjunction with the hybrid Monte Carlo (HMC) scheme to improve samples generated by standard particle filters. Parallel marginalization is an efficient Markov chain Monte Carlo (MCMC) strategy that uses lower dimensional approximate marginal distributions of the target distribution to accelerate equilibration. As a validation the algorithm is tested on a 2516 dimensional, bimodal, stochastic model motivated by the Kuroshio current that runs along the Japanese coast. The results of this test indicate that the method is an attractive alternative for problems that require the generality of a particle filter but have been inaccessible due to the limitations of standard particle filtering strategies.

Original languageEnglish (US)
Pages (from-to)4312-4331
Number of pages20
JournalJournal of Computational Physics
Volume228
Issue number12
DOIs
StatePublished - Jul 1 2009

Fingerprint

ocean currents
Ocean currents
sampling
Sampling
Stochastic models
filters
Markov processes
Coastal zones
Markov chains
coasts

Keywords

  • Hybrid Monte Carlo
  • Kuroshio
  • Parallel marginalization
  • Particle filter
  • Path sampling
  • Weather prediction

ASJC Scopus subject areas

  • Physics and Astronomy (miscellaneous)
  • Computer Science Applications

Cite this

Particle filtering with path sampling and an application to a bimodal ocean current model. / Weare, Jonathan.

In: Journal of Computational Physics, Vol. 228, No. 12, 01.07.2009, p. 4312-4331.

Research output: Contribution to journalArticle

@article{de03f0a3d7854149ba8a5cc8ecc25b1f,
title = "Particle filtering with path sampling and an application to a bimodal ocean current model",
abstract = "This paper introduces a recursive particle filtering algorithm designed to filter high dimensional systems with complicated non-linear and non-Gaussian effects. The method incorporates a parallel marginalization (PMMC) step in conjunction with the hybrid Monte Carlo (HMC) scheme to improve samples generated by standard particle filters. Parallel marginalization is an efficient Markov chain Monte Carlo (MCMC) strategy that uses lower dimensional approximate marginal distributions of the target distribution to accelerate equilibration. As a validation the algorithm is tested on a 2516 dimensional, bimodal, stochastic model motivated by the Kuroshio current that runs along the Japanese coast. The results of this test indicate that the method is an attractive alternative for problems that require the generality of a particle filter but have been inaccessible due to the limitations of standard particle filtering strategies.",
keywords = "Hybrid Monte Carlo, Kuroshio, Parallel marginalization, Particle filter, Path sampling, Weather prediction",
author = "Jonathan Weare",
year = "2009",
month = "7",
day = "1",
doi = "10.1016/j.jcp.2009.02.033",
language = "English (US)",
volume = "228",
pages = "4312--4331",
journal = "Journal of Computational Physics",
issn = "0021-9991",
publisher = "Academic Press Inc.",
number = "12",

}

TY - JOUR

T1 - Particle filtering with path sampling and an application to a bimodal ocean current model

AU - Weare, Jonathan

PY - 2009/7/1

Y1 - 2009/7/1

N2 - This paper introduces a recursive particle filtering algorithm designed to filter high dimensional systems with complicated non-linear and non-Gaussian effects. The method incorporates a parallel marginalization (PMMC) step in conjunction with the hybrid Monte Carlo (HMC) scheme to improve samples generated by standard particle filters. Parallel marginalization is an efficient Markov chain Monte Carlo (MCMC) strategy that uses lower dimensional approximate marginal distributions of the target distribution to accelerate equilibration. As a validation the algorithm is tested on a 2516 dimensional, bimodal, stochastic model motivated by the Kuroshio current that runs along the Japanese coast. The results of this test indicate that the method is an attractive alternative for problems that require the generality of a particle filter but have been inaccessible due to the limitations of standard particle filtering strategies.

AB - This paper introduces a recursive particle filtering algorithm designed to filter high dimensional systems with complicated non-linear and non-Gaussian effects. The method incorporates a parallel marginalization (PMMC) step in conjunction with the hybrid Monte Carlo (HMC) scheme to improve samples generated by standard particle filters. Parallel marginalization is an efficient Markov chain Monte Carlo (MCMC) strategy that uses lower dimensional approximate marginal distributions of the target distribution to accelerate equilibration. As a validation the algorithm is tested on a 2516 dimensional, bimodal, stochastic model motivated by the Kuroshio current that runs along the Japanese coast. The results of this test indicate that the method is an attractive alternative for problems that require the generality of a particle filter but have been inaccessible due to the limitations of standard particle filtering strategies.

KW - Hybrid Monte Carlo

KW - Kuroshio

KW - Parallel marginalization

KW - Particle filter

KW - Path sampling

KW - Weather prediction

UR - http://www.scopus.com/inward/record.url?scp=67349260731&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=67349260731&partnerID=8YFLogxK

U2 - 10.1016/j.jcp.2009.02.033

DO - 10.1016/j.jcp.2009.02.033

M3 - Article

VL - 228

SP - 4312

EP - 4331

JO - Journal of Computational Physics

JF - Journal of Computational Physics

SN - 0021-9991

IS - 12

ER -