Quaternion-valued robust adaptive beamforming based on principal eigenspace projection

Based on the principal eigenspace projection in the quaternion domain, a robust adaptive beamforming scheme was proposed with a dual-parallel array of spatially stretched tripole antennas, to tackle the performance degradation of quaternion-based adaptive beamformers in the presence of model mismatch errors. Compared with the quaternion-based worst-case constrained beamformer, the presented method does not need convex optimization and user-parameter selection. Numerical simulations show that the proposed method can tackle the signal self-nulling problem effectively, and significantly outperforms the quaternion-based worst-case constrained beamformer in the case of high SNRs and small sample sizes with reduced computational complexity.