Robust Capon estimator for the nominal DOAs and angular spreads of distributed sources

A novel and low complexity estimator based on Capon´s minimum variance beamforming method is proposed for the estimation of the nominal (centre) directions of arrival (DOA) and angular spreads of distributed sources. A more reliable array response vector instead of the standard steering vector is put forward according to the incoherently distributed (ID) source model, by which the spread Capon spectrum is obtained. Considering the model mismatch between the new steering vector and the true one, we estimate the parameters not by making the trace of the Capon spectrum matrix minimization but, making the corresponding Frobenius norm minimization. Only a two-dimensional search method is required to optimize the estimates of the nominal angles and angular spreads, so the proposed method is computationally efficient. Numerical results demonstrated its excellent performance in the low SNR case and verified its robustness against some other methods.