Self-steered Robust Capon Beamformer Against Large Look Direction Error

A self-steered robust Capon beamformer is proposed to address the problem of significant performance degradation caused by the large look direction error. The basic idea behind the proposed method is to reconstruct the interference-plus-noise covariance (INC) matrix precisely. First, the noise covariance matrix is estimated by the eigen-decomposition. Then, the interference covariance matrix is obtained by integrating the signal power outside the region-of-interest (ROI) using the maximum entropy power spectrum (MEPS) estimator. Simulation results show that the proposed beamformer's main beam can be self-steered to the real direction-of-arrival (DOA) of the signal, thus can effectively suppress the large look direction error and achieve the performance close to the theoretical optimal value.