Abstract: The development of diagonal loading beamforming technique is reviewed and some methods of computing diagonal loading factor are presented.Then,by simulation examples,the infinite-random-matrix-based beamforming presented by Xavier is analyzed.The result shows that the sell-nulling phenomenon happens in the scenario with steering vector errors to degrade the performance severely.Due to the orth-ogonality between the steering vector of the signal of interest(SOI) and the noise space,an improved method elicited by the eigenspace-based beamforming(ESB) is proposed,in which the direction of SOI is first calibrated.This method can improve the robustness of the beamformer in the case of Direction-of-Arrival(DOA) uncertainties.The result of the numerical simulation shows the effectiveness of this method when the SNR is not great.