Widely Linear Constrained Complex-Valued Least M-estimation Algorithm for Adaptive Beamforming

In this paper, a widely-linear constrained complex-valued least mean M-estimate (WL-CCLMM) algorithm is derived for processing impulsive noise and complex-valued signal in adaptive beamforming. In the proposed algorithm, a widely-linear structure is used to improve steady-state performance by making use of the second-order (SO) non-circular property of the input random vector. Weight constraints are introduced to prevent distortion in desired signal directions and nulls in interference signal directions. We also adopt the M-estimator function to decrease the influence of outliers caused by impulsive noise. Experimental results verify that the WL-CCLMM algorithm realizes lower steady-state performance and better beam pattern effect in adaptive beamforming with different noise environment, compared to other previous robust algorithms.