Robust adaptive beamforming based on reconstruction of the covariance matrix and the estimation of steering vector

In the practical application, the error between the assumed steering vector of the desired signal and the real one causes the sharp degrading of the beamforming performance, especially when the desired signal power is strong. To solve this problem, a novel robust algorithm is proposed. The interference-plus-noise covariance matrix is sparse when the number of the signal source is less than the array element number. At first, by using this sparse property, the interference-plus-noise covariance matrix can be reconstructed, from which a subspace orthogonal to the steering vectors of interference signals can be obtained. Then, a mixed signal only containing the desired signal and the noise can be gained by mapping the received data through to this orthogonal subspace. The real steering vector of the desired signal can be estimated based on maximizing the out power of the mixed signal. The final beamforming weight can be constructed by using the orthogonal subspace and the real steering vector of the desired signal. Simulation results demonstrate the efficiency of the proposed algorithm in the case of the mismatch of the steering vector of the desired signal, the strong desired signal and the low snapshot number.