WIT Press

An Enhanced Methodology For Diagnosing And Predicting Air Pollution States

Price

Free (open access)

Paper DOI

10.2495/AIR940121

Volume

3

Pages

8

Published

1994

Size

696 kb

Author(s)

M. Nonaka & N.H. Thomas

Abstract

An enhanced methodology for diagnosing and predicting air pollution states M. Nonaka", N.H. Thomas* "Department of Geosystem Engineering, The University of Tokyo, Tokyo 113, Japan School of Chemical Engineering, The University of Birmingham, Birmingham B15 2TT, UK ABSTRACT The switching mode enhanced extended Kalman(SEEK) algorithm has been developed to estimate the states of strongly nonlinear and noisy dynami- cal systems such as contaminants dispersion processes in the atmosphere or ocean. The switching parameter in the observation matrix is optimally selected so as to ensure the convergence rate of the estimation error covari- an ce matrix is maximised. As a model illustration here the SEEK algorithm has been applied to estimate the state variables of the Lorenz system deeply embedded within the Gaussian white noise. The SEEK filter has returned accurate state estimates of this chaotic system, whilst the conventional ex- tended K aim an filter has failed to e

Keywords