Bayesian inference for source determination with applications to a complex urban environment

Keats, Andrew; Yee, Eugene; Lien, Fue‐Sang

doi:10.1016/j.atmosenv.2006.08.044

Cited by 262 publications

(192 citation statements)

References 21 publications

Supporting

Mentioning

190

Contrasting

Unclassified

Order By: Relevance

“…However, as it is computationally expensive, an efficient sampling technique is required to approximate the posterior distribution. Within the literature on STE, several techniques have been used: i) Markov chain Monte Carlo (MCMC) (Keats et al (2007); Senocak et al (2008)); ii) sequential Monte Carlo (SMC) (Johannesson et al (2005)); and iii) differential evolution Monte Carlo (DEMC) (Robins et al (2009)). In this work, we use the Metropolis-Hastings MCMC algorithm (Hastings (1970)).…”

Section: Inferencementioning

confidence: 99%

“…The most popular approach to STE features a static network of concentration sensors spread over a large region on the ground. Source estimation is then carried out using optimization (Long et al (2010); Thomson et al (2007)) or Bayesian inference algorithms (Keats et al (2007); Senocak et al (2008)) where inferred source parameters are run in a forward ATD model to generate predicted concentrations that are then compared with the data using a cost or likelihood function. A recent study by Platt and Deriggi (2010) based on data from the FFT07 experiment demonstrated some of the limitations of theses approaches when applied to real data.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Adaptive Bayesian Sensor Motion Planning for Hazardous Source Term Reconstruction

Hutchinson

Wen-hua

2017

IFAC-PapersOnLine

View full text Add to dashboard Cite

There has been a strong interest in emergency planning in response to an attack or accidental release of harmful chemical, biological, radiological or nuclear substances. Under such circumstances, it is of paramount importance to determine the location and release rate of the hazardous source to forecast the future harm it may cause and employ methods to minimize the disturbance. In this paper, a sensor data collection strategy is proposed whereby an autonomous mobile sensor is guided to address such a problem with a high degree of accuracy and in a short amount of time. First, the parameters of the release source are estimated using the Markov chain Monte Carlo sampling approach. The most informative manoeuvre from the set of possible choices is then selected using the concept of maximum entropy sampling. Numerical simulations demonstrate the superior performance of the proposed algorithm compared to traditional approaches in terms of estimation accuracy and the number of measurements required.

show abstract

Section: Inferencementioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Adaptive Bayesian Sensor Motion Planning for Hazardous Source Term Reconstruction

Hutchinson

Wen-hua

2017

IFAC-PapersOnLine

View full text Add to dashboard Cite

show abstract

“…The preferred alternative is the use of Bayesian techniques; they result in the posterior probability density function (PDF) of the source parameter vector, thereby providing an uncertainty measure to any point estimate derived from it. Most Bayesian methods for source estimation are based on Markov chain Monte Carlo (MCMC) technique, assuming either Gaussian or log-Gausiian likelihood function of measurements [5,6,7,8]. Recently, a likelihood-free Bayesian method for source localisation was proposed in [9].…”

Section: Introductionmentioning

confidence: 99%

Localisation of a source of hazardous substance dispersion using binary measurements

Ristić

Gunatilaka

Gailis

2016

Atmospheric Environment

View full text Add to dashboard Cite

Using the measurements collected at a number of known locations by a moving binary sensor, characterised by an unknown threshold, the problem is to estimate the parameters of a biochemical source, continuously releasing material into the atmosphere. The solution is formulated in the Bayesian framework using a dispersion model of Poisson distributed particle encounters in a turbulent flow. The method is implemented using the importance sampling technique and successfully validated with three experimental datasets under different wind conditions.

show abstract

“…Several recent event reconstruction studies have favored the Bayesian inference approach over the optimization approach as it offers several advantages (Johannesson et al, 2004;Chow et al, 2006;Keats et al, 2007). The main distinguishing feature of the Bayesian inference method is that it estimates probability distributions for parameters of interest and quantifies the uncertainty in the estimated parameter, whereas an optimization method provides point estimates for the parameters of interest through maximizing or minimizing an objective function.…”

Section: Introductionmentioning

confidence: 99%

“…The results have shown that significant gains in computational time can be obtained by adopting the new scheme over direct sampling. Keats et al (2007) considered a source-receptor relationship within the Bayesian inference method to reduce the overall computation time for source determination. An adjoint equation for the contaminant concentration was solved for that purpose.…”

Section: Introductionmentioning

confidence: 99%

Stochastic event reconstruction of atmospheric contaminant dispersion using Bayesian inference

Şenocak

Hengartner

Short

et al. 2008

Atmospheric Environment

View full text Add to dashboard Cite

This is an author-produced, peer-reviewed version of this article. AbstractEnvironmental sensors have been deployed in various cities for early detection of contaminant releases into the atmosphere. Event reconstruction and improved dispersion modeling capabilities are needed to estimate the extent of contamination, which is required to implement effective strategies in emergency management. To this end, a stochastic event reconstruction capability that can process information from an environmental sensor network is developed. A probability model is proposed to take into account both zero and non-zero concentration measurements that can be available from a sensor network because of a sensor's specified limit of detection. The inference is based on the Bayesian paradigm with Markov chain Monte Carlo (MCMC) sampling. Fast-running Gaussian plume dispersion models are adopted as the forward model in the Bayesian inference approach to achieve rapid-response event reconstructions. The Gaussian plume model is substantially enhanced by introducing stochastic parameters in its turbulent diffusion parameterizations and estimating them within the Bayesian inference framework. Additionally, parameters of the likelihood function are estimated in a principled way using data and prior probabilities to avoid tuning in the overall method, The event reconstruction method is successfully validated for both real and synthetic dispersion problems, and posterior distributions of the model parameters are used to generate probabilistic plume envelopes with specified confidence levels to aid emergency decisions.Key words: Bayesian Statistics, Event Reconstruction, Source Characterization, Gaussian Plume Models, Markov chain Monte Carlo (MCMC) Preprint submitted to Atmospheric Environment 27 April 2008This is an author-produced, peer-reviewed version of this article. The final, definitive version of this document can be found online at Atmospheric Environment,

show abstract

Bayesian inference for source determination with applications to a complex urban environment

Cited by 262 publications

References 21 publications

Adaptive Bayesian Sensor Motion Planning for Hazardous Source Term Reconstruction

Adaptive Bayesian Sensor Motion Planning for Hazardous Source Term Reconstruction

Localisation of a source of hazardous substance dispersion using binary measurements

Stochastic event reconstruction of atmospheric contaminant dispersion using Bayesian inference

Contact Info

Product

Resources

About