Robust adaptive beamforming using modified constant modulus algorithms

Abstract This paper addresses the self-nulling phenomenon also known as the self-cancellation in adaptive beamformers. Optimum beamforming requires knowledge of the desired signal characteristics, either its statistics, its direction-of-arrival, or its response vector. Inaccuracies in the required information lead the beamformer to attenuate the desired signal as if it were interference. Self-nulling is caused by the desired signal having large power (high SNR) relative to the interference signal in case of the minimum variance distortion less response beamformer, and low power desired signal in the case of the constant modulus algorithm (CMA) beamformer, which leads the beamformer to suppress the desired signal and lock onto the interference signal. The least-square constant modulus algorithm is a prominent blind adaptive beamforming algorithm. We propose two CMA-based algorithms which exploit the constant modularity as well as power or DOA of the desired signal to avoid self-nulling in beamforming. Simulations results verify the effectiveness of the proposed algorithms.