|
Abstract:
|
The threat of Chemical, Biological, Radiological and Nuclear (CBRN) attack is a pertinent feature of the modern battlefield. Rapid detection, assessment and early response to CBRN releases could dramatically reduce the extent of human exposure, help maintain operational tempo and minimize the cost of the subsequent clean up. To this end, by characterizing the plume through time, either directly or via source term estimation and a dispersion model, prediction of the dispersion of the contaminant can be made. A methodology is developed for making inference about source term parameters from point sensor readings where results must be obtained in a very short timescale. An SMC approach is used where there are no state evolution equations, the forward model is highly non-linear and the likelihoods are non-Gaussian. Further, an approximate Bayesian computation algorithm (ABC) is applied.
|
