Mixed source localization and gain-phase perturbation calibration in partly calibrated symmetric uniform linear arrays

摘要

This paper is concerned with mixed far-field and near-field source localization in the presence of gain-phase perturbations. Under some mild assumptions, we propose a new mixed source localization algorithm with the partly calibrated symmetric uniform linear array. By constructing a special second-order statistical vector, the DOAs and powers of all sources are first estimated by the modified sparse total least square (M-STLS) algorithm after compensating gain errors. Based on the estimated DOAs, the phase errors are successively obtained by the least squares criterion and a discriminant function formed by using the MUSIC null spectrum property. Finally, the mixed sources are classified and the range of near-field sources are achieved via one-dimensional spectral search. Meanwhile, the stochastic Cramer-Rao bound (CRB) for the considered problem is also given. The proposed algorithm can lead to a good mixed source classification and localization result. Numerical simulations validate the effectiveness of the proposed algorithm. (C) 2019 Elsevier B.V. All rights reserved.