Noncircular signal regularized-beamforming for defective acoustic vector arrays

Investigates the noncircularity-exploration of noncircular signals in adaptive beamforming with defective acoustic vector-arrays. Constructs the conjugate signal plus interference subspace that is immune to colored noise by taking into account nonvanishing moments of signals, then soft projection of adaptive weights is performed onto this subspace with a regularization scheme, to enlarge the smaller eigenvalues of the correlation matrix of the array output. The effect of perturbation of these smaller eigenvalues can thus be greatly alleviated. Simulation results show that, in the presence of colored noise and/or look-direction steering error, the proposed method has a stronger capability on remedying beam pattern distortion and eliminating signal cancellation, when compared with the conventional method. Using regularization scheme to fuse the noncircularity property of signals can remarkably improve the adaptive beamforming performance with colored noise and/or steering error present.