![]() | Only 14 pages are availabe for public view |
Abstract The thesis aims to improve the performance of adaptive antenna arrays specifically the GSC beamformer using wavelet transform based techniques. Three techniques are discussed, The tlrst one is based on the use of the high pass wavelet filters in the blocking matrix of the GSC beamformer to perform spatial filtering of the received signal, thus blocking the desired signal from entering the adaptive portion of the beamformer and giving an accurate estimation of the interference signal, this technique has been proposed before in [13] for the narrowband GSC under the name of the wavelet based GSc. In this thesis we extend the wavelet based sidelobe canceller to the more general wideband wavelet packet based GSC beamformer and show through simulation results the improvement in the interference rejection capability when using’this novel beamformer especially for jammers in the vicinity of the steer direction, furthermore, it has the effect of decreasing the number of adaptive weights required for interference estimation through a proposed subband selection scheme. The second one is based on the use of wavelet transform domain LMS algorithms in the adaptive portion of the beamformer, in order to decrease the spread in the eigenvalues of the correlation matrix of the input data which causes a faster convergence rate of the adaptive mformer. The third uses the wavelet transform domain self-orthogonalizing LMS algorithm in he adaptive portion of the beamformer. Aiming here at obtaining a sparse estimate of the rrelation matrix of the input data, which decreases the number of operations required to for he inversion of this matrix, without much affecting the convergence rate or the steady state lution of the algorithm. |