Title:Low-Complexity On-Grid Channel Estimation for Partially-Connected Hybrid XL-MIMO

Abstract:This paper addresses the challenge of channel estimation in extremely large-scale multiple-input multiple-output (XL-MIMO) systems, pivotal for the advancement of 6G communications. XL-MIMO systems, characterized by their vast antenna arrays, necessitate accurate channel state information (CSI) to leverage high spatial multiplexing and beamforming gains. However, conventional channel estimation methods for near-field XL-MIMO encounter significant computational complexity due to the exceedingly high parameter quantization levels needed for estimating the parametric near-field channel. To address this, we propose a low-complexity two-stage on-grid channel estimation algorithm designed for near-field XL-MIMO systems. The first stage focuses on estimating the LoS channel component while treating the NLoS paths as interference. This estimation is accomplished through an alternating subarray-wise array gain maximization (ASAGM) approach based on the piecewise outer product model (SOPM). In the second stage, we estimate the NLoS channel component by utilizing the sensing matrix refinement-based orthogonal matching pursuit (SMR-OMP) algorithm. This approach helps reduce the high computational complexity associated with large-dimensional joint sensing matrices. Simulation results demonstrate the effectiveness of our proposed low-complexity method, showcasing its significant superiority over existing near-field XL-MIMO channel estimation techniques, particularly in intermediate and high SNR regimes, and in practical scenarios involving arbitrary array placements.