Low-complexity blind beamforming based on cyclostationarity

A blind adaptive beamformer with real-valued weights for cyclostationary signals is proposed by introducing a preprocessing stage. It is proved that the optimum weight vector of the proposed method is real-valued. Thus we can discard the imaginary part of the weight vector during the update process, which reduces the computational complexity of the algorithm. Moreover, since the optimum solution is real-valued, a result closer to the optimum one is reached, leading to a much increased convergence rate.