Low‐complexity variable loading for robust adaptive beamforming

Although the majority of the existing robust adaptive beamformers focus on how to choose an identical loading factor for all eigenvalues, relatively few investigations on variable loading (VL) have been conducted. A low‐complexity VL beamformer is proposed in which the beamformer weight vector is deliberately prevented from converging to the noise components. Then inspired by the traditional identical loading method, the loading factor can be set in an ad hoc manner. Numerical results demonstrate the superior performance of the proposed beamformer relative to other existing approaches such as ‘identical diagonal loading’ and ‘robust Capon’ beamformers.