Unity And Common Information Fusion White Noise Deconvolution Estimators

Multi sensor information fusion is a new technology to process multidimensionalinformation synthetically. It has been applied in the field of information acquisition andprocessing widely and has become an active research focus in current informationdomain. The theory of white noise estimation is an important researching aspect. it hasan important applied background in oil seismic exploration. The principle of oil seismicexploration is to utilize information which is provided by the series of coefficients ofreflection which is produced by seismic wave in oil layer through explosion ofunderground explosive to judge the size of geometrical shape of oil field. Because theseries of coefficients of reflection can be described by Bernoulli-Gaussian white noise,the white noise estimation becomes the key problem in oil seismic exploration. Mendelhad solved the problem using Kalman filtering method and proposed the in-put whitenoise estimator of system, which is also called white noise deconvolution filter.For the multi sensor linear discrete time-varying stochastic control systems withthe different local dynamic models, using the Kalman filtering method, tinder theoptimal fusion criterion weighted by scalars, the time-varing optimal information fusionwhite noise deconvolution estimators are presented, and for the correspondingtime-invariant systems, the non-steady-state optimal information fusion white noisedeconvolution estimators and steady-state information fusion white noise deconvolutionestimators are also presented. They can be used to handle the white noise deconvolutionfused filtering, smoothing and prediction problems in a unified framework. Especially,they can be applied to solve the estimation problem of system with color observationnoise. In order to compute the optimal weights, the formula of computing thecross-covariances among local estimation errors of input white noise is given. A lot ofMonte Carlo simulation examples for Bernoulli-Gaussian white noise are used to show the above theories effectiveness and correctness.