Two-dimensional wavelet-based generalized sidelobe canceller with subband selection scheme

We present two approaches to the design of two-dimensional (2-D) blocking matrices of the generalized sidelobe canceller (GSC). The first approach is an extension of the wavelet-based partially adaptive beamformer considered in . It can be shown that this 2-D blocking matrix with the regularity P nulls out the first (P-1) order Taylor's series expansion of the desired signal. Hence, the derivative constraints can be used to establish a flatter mainlobe response. The second method uses the subband decomposition characteristic of these wavelets. Only the principal subband component (PSC) of the decomposed data is kept for adaptive process. As such, the overall computational complexity is substantially reduced. The resulting adaptive beamformer is less sensitive to steering angle errors.