Self-tuning Information Fusion Wiener Filter for Multisensor Multichannel AR Signals

Abstract

For the multisensor multichannel autoregressive (AR) signals with unknown model parameters and noise variances, the estimators of model parameters and noise variances can be obtained based on the multidimensional recursive extended least squares (RELS) algorithm and the correlation method. Furthermore, a self-tuning information fusion Wiener filter is presented based on the modern time series analysis method by substituting the estimators for the true values. One simulation example shows the consistence of the estimators of the model parameters and noise variances and the convergence of the self-tuning Wiener filter.