\documentclass[11pt]{article}
\begin{document}
\title{{Monte Carlo Filtering Algorithm for Nonlinear Stochastic
Environmental Model} \\ {\em (Abstract)}} \author{ V. P. Jilkov and
Tz. A. Semerdjiev \\ Laboratory for Parallel and Distributed
Processing \\ Bulgarian Academy of Sciences \\ tel: (+3592) 731 498,
fax: (+3592) 707 273, e-mail: signal@acad.bg} \maketitle
\thispagestyle{empty}

The work presents development and evaluation of an alternative
Monte Carlo simulation filtering algorithm for application to
environmental (atmoshperic, ocean etc.) data assimilation.

Most of existing and proposed data assimilation schemes are designed
on the assumptions that the estimate is the mean of Gaussian
distribution and near-linearity of process dynamics, which do not
match real processes adequately. Few nonlinear data assimilation
algorithms are known, which are based on the classical nonlinear
filtering theory, but their implementation is hampered by many
restrictions and practical difficulties.

A recently developed approach to nonlinear/non-Gaussian state
estimation - the {\em bootstrap filtering} is adopted and applied in
our work. A bootstrap data assimilation algorithm is designed and
applied to the stochastically forced {\em double-well model}. The
algorithm implements recursive Bayesian filtering and does not
require the limiting assumptions and solving of the  Fokker-Planck
stochastic differential equations.

By means of Monte Carlo simulation experiments, the performance of the
proposed algorithm is evaluated and compared with known algorithms -
Extended Kalman filter and Fokker-Planck equation based algorithm.

\end{document}