Framework for Practical Characterization of Spatial Non-stationary ELAA Propagation Channels

In this paper, a complete framework for the practical evaluation of extremely large antenna array (ELAA) channels is proposed. Based on the empirical observed spatial non-stationary characteristics (SNC), namely the variable magnitudes (VMs) of the multipath components (MPCs) with the antenna element position, a new channel modeling method considering the near-field effect and appropriate to parametrize the SNC was proposed. The model characterizes the VM of each MPC observed for the different sub-channels of an ELAA by Rician and Rayleigh distributions, which violates the stationarity assumptions of traditional parameter estimation methods, such as the space-alternating generalized expectation-maximization (SAGE). In order to cope with this, a novel estimation algorithm, namely the Spectrum-Analysis-Aided Multipath-Component Extraction (SA-MPCE), was proposed as well. Finally, in order to enable the practical and time-affordable empirical characterization of ELAA channels, a switch-based channel sounder, as well as a fast calibration method, were designed. In order to show the applicability of the proposals, an exemplary measurement-based channel model was provided. The presented methodologies provide a complete architecture for the practical research on ELAA for future communication systems, including the data acquisition, parameters estimation and channel modeling.