It is well know that the performance of the minimum\udvariance distortionless response (MVDR) beamformer\udis very sensitive to steering vector mismatch because of the\udsignal cancellation. Such mismatches can occur as a result of\uddirection-of arrival (DOA) errors, local scattering, near-far spatial\udsignature mismatch, waveform distortion, source spreading,\udimperfectly calibrated arrays and distorted antenna shape.\udIn this paper, an adaptive beamformer which is robust against\udthe DOA mismatch is proposed. This method imposes two\udquadratic constraints such that the magnitude responses of two\udsteering vectors exceed unity. Then a diagonal loading method\udis used to force the magnitude responses within a range of the\udarrival angles to exceed unity. The diagonal loading factor can\udbe computed systematically by a proposed algorithm. Numerical\udexamples show that this method has a significantly better SINR\udperformance compared to previously published methods (including\udthe MVDR), and a complexity comparable to the standard\udMVDR beamformer.
展开▼
