Wavelet-based communication channel modeling and identification
Milos I. Doroslovacki, Ljubiša Stanković · Proceedings of SPIE, the International Society for Optical Engineering/Proceedings of SPIE · 2002
The theory of modeling linear communication channels by wavelets is reviewed. It is shown that wavelet-based channel identification can converge fast and be a flexible tool for dealing with many different signal propagation situations. Moreover, wavelet-based identification can provide robustness against narrow band noise and/or impulse noise. Possibilities for sparse channel representations are emphasized and illustrated by real life examples. A wavelet packet decomposition that maximizes the convergence speed of the LMS algorithm used in channel identification is discussed. Comparisons with the DFT, DCT, and other transforms are given.