Characterizing and Utilizing Near-Field Spatial Correlation for XL-MIMO Communication

This paper investigates the near-field spatial correlation (SC) for extremely large-scale multiple-input multiple-output (XL-MIMO) communications, where the positions of scatterers and/or users may be in the radiative near-field region due to the large size of the antenna array. Therefore, the conventional far-field uniform plane wave (UPW) assumption is no longer valid. By discarding the UPW assumption, we consider the generic non-uniform spherical wave (NUSW) model for accurate characterization of signal amplitude and phase variations on different antenna elements. A novel expression is derived for the near-field SC in terms of the scatterer distribution, which generalizes the conventional far-field SC. It is revealed that the developed near-field SC is determined by the power location spectrum (PLS), which is characterized not only by the scatterers' directions but also by their distances from the array, while the conventional far-field model is only dependent on the power angular spectrum (PAS). In addition, the near-field SC no longer exhibits spatial wide-sense stationarity (SWSS), since the SC coefficient between each pair of transmit-receive antenna components is determined by their specific positions, not just by their relative locations. Furthermore, the developed near-field SC is utilized to obtain the optimal transmission strategy to maximize the ergodic spectral efficiency based on statistical channel state information (CSI). Besides, we consider the specific multi-ring scattering model for scatterer distribution, where semi-closed expressions for the near-field SC can be obtained. Simulation results are given to validate the developed near-field SC for XL-MIMO communications.