Coprime array-based robust beamforming using covariance matrix reconstruction technique

A novel robust adaptive beamforming algorithm based on a coprime array is proposed in this study. First, the authors exploit the virtual array structure of the coprime coarray to construct two subspaces. The first subspace is obtained from the covariance matrix of the virtual uniform linear array, whereas the second one is obtained by integrating a spatial spectrum over each angular sector of the signal. Then, by using a closed-form expression of the projection-based method, the steering vectors (SVs) of the signals are estimated from the intersection of the two subspaces. Moreover, according to the covariance fitting theory, the power associated with the desired signal, interference signals, and noise is obtained from the virtual sample covariance matrix. In addition, to reconstruct the interference-plus-noise covariance matrix (INCM), the maximum correlation principle is used to transfer the virtual SVs of the signals to real ones corresponding to a physical array. Finally, with the estimated SV of the desired signal and INCM, the proposed robust algorithm is devised. Simulation results demonstrate the superiority and effectiveness of the proposed algorithm.