Parameter estimation of a continuous chemical plume source

The problem is estimation of the strength (emission rate) and the location of a chemical source from which a contaminant is released continuously into the atmosphere. The concentration of the contaminant is measured at regular intervals by a network of spatially distributed chemical sensors. The transport of the contaminant is modeled by turbulent diffusion and fluctuating plume concentration. The source parameter estimation is solved in the sequential Bayesian framework, with the posterior expectation approximated using Monte Carlo integration. In order to deal with the vague prior, importance sampling is implemented using the progressive correction technique. The paper presents numerical analysis of statistical performance of the proposed algorithms for different sensor configurations.