Robust blind beamforming using neural network

Many blind beamforming algorithms, such as constrained cyclic adaptive beamforming (C-CAB), use cyclostationarity to estimate the steering vector and adaptively obtain the linearly constrained minimum variance (LCMV) optimum solution. However, LCMV methods are sensitive to the mismatch caused by the uncalibration array or estimate error. After discussion of this mismatch, a robust blind beamforming algorithm is presented. Implemented as a neural network, this algorithm reduces the computational complexity for real-time use. Computer simulations verify the analysis