WAVELET-BASED SHIFT-INVARIANT NOISE MODEL

Abstract

This paper proposes the use of shift-invariant dual-tree discrete wavelet transform (DT DWT) noise model to the problem of a multipath signal detection in underwater environment. The design procedure for this adaptive model of the background noise is described. When an input signal is shifted, as in a multipath environment, this shift-invariant DT DWT can generate multi-resolution subspaces. Each of them keeps more coefficient energy of this signal than DWT’s. The result of computer simulations shows that this method can improve the performance of a multipath signal detection, by reducing the false alarm probability in a range of different signal-tonoise (SNR) ratios.