Abstract
In this paper, we propose an efficient near-field channel reconstruction and user localization scheme for extremely large-scale antenna array (ELAA) systems. Due to the non-negligible near-field effect in ELAA systems, a more realistic near-field multipath channel model, which incorporates the unequal path loss and the phase deviations across antennas and models line-of-sight (LoS), reflection, and scattering, is considered. A subarray hybrid beamforming architecture is further employed to reduce the cost of using ELAA. Based on the sparsity of the near-field channel in the joint angle-distance domain, a near-field Newtonized orthogonal matching pursuit algorithm is proposed to estimate the multipath parameters. Reconstruction of the near-field channel and positioning of the user can be achieved based on the estimated parameters. Our numerical results verify that the reconstructed channel is very close to the real near-field channel, and the user localization has high accuracy whether a LoS component exists or not, validating the effectiveness and reliability of the proposed scheme.