Filtering of colored noise for signal enhancement

Abstract

This paper suggests an enhancement approach for signal corrupted by additive colored noise signal. The well known FIR structure filter is adopted in order to obtain the noise-suppressed estimate of the desired signal. It is shown that the suggested approach has the quick estimation ability for desired signal. It is also shown that the estimate of the desired signal is separated from the additive colored noise signal when the additive colored noise signal is nearly constant on the window. In addition, when the additive colored noise signal itself is treated as an additional desired signal that should be estimated, its estimate is shown to be separated from the state term for the original desired signal. Via numerical simulations on a military signal, the performance of the suggested approach is evaluated by the comparison with that of the existing Kalman filtering approach. © Springer-Verlag Berlin Heidelberg 2004.