In this work, we investigate the channel estimation (CE) problem for extremely large-scale multiple-input-multiple-output (XL-MIMO) systems, considering both the spherical wavefront effect and spatial non-stationarity (SnS). Unlike existing non-stationary CE methods that rely on the statistical characteristics of channels in the spatial or temporal domain, our approach seeks to leverage sparsity in both the spatial and wavenumber domains simultaneously to achieve an accurate estimation.To this end, we introduce a two-stage visibility region (VR) detection and CE framework. Specifically, in the first stage, the belief regarding the visibility of antennas is obtained through a structured message passing (MP) scheme, which fully exploits the block sparse structure of the antenna-domain channel. In the second stage, using the obtained VR information and wavenumber-domain sparsity, we accurately estimate the SnS channel employing the belief-based orthogonal matching pursuit (BB-OMP) method. Simulations demonstrate that the proposed algorithms lead to a significant enhancement in VR detection and CE accuracy, especially in low signal-to-noise ratio (SNR) scenarios.

翻译：本文针对超大规模多输入多输出（XL-MIMO）系统的信道估计问题展开研究，同时考虑球面波前效应和空间非平稳性（SnS）。与现有依赖信道在空间或时域统计特性的非平稳信道估计方法不同，本文方法旨在同时利用空间域和波数域中的稀疏性实现精确估计。为此，我们提出了一种两阶段可见区域（VR）检测与信道估计框架。具体而言，在第一阶段，通过结构化消息传递（MP）方案获取天线可见性的置信度，该方案充分利用天线域信道的块稀疏结构。在第二阶段，利用所获取的VR信息与波数域稀疏性，采用基于置信度的正交匹配追踪（BB-OMP）方法对SnS信道进行精确估计。仿真结果表明，所提算法在VR检测与信道估计精度上均有显著提升，尤其在低信噪比（SNR）场景下表现突出。