发表机构
City University of Hong Kong; The University of Hong Kong（香港城市大学; 香港大学）
机构由 AI 辅助整理，请以论文原文为准。AI 中文总结
针对XL-MIMO空间非平稳信道估计，提出联合角度-子阵列码本与Gram注意力学习框架，将问题转化为稀疏支撑检测，实现最高F1分数和最低NMSE。
AI 中文摘要
随着下一代无线系统向高频段迁移，超大规模多输入多输出（XL-MIMO）对于对抗严重的路径损耗不可或缺。然而，沿着超大规模阵列孔径（ELAA）接收的路径能量可能表现出空间非平稳性，这使得传统的离散傅里叶变换（DFT）码本因其全阵列角度表示而不足以用于信道估计（CE）。为解决这一失配问题，我们提出了一种新颖的联合角度-子阵列（JAS）码本，其中每个码字对应一个角度方向和一个特定子阵列。空间非平稳信道估计因此被转化为JAS域中的结构化稀疏支撑检测。然而，对于传统的基于模型的方法，可靠的支撑恢复仍然具有挑战性。因此，我们开发了一种信号处理与深度学习协同设计框架，其中数据驱动的JAS并行支撑检测之后是基于最小二乘（LS）的信道重构。每个JAS网格的匹配滤波响应被嵌入到高维特征向量中。然后，我们提出了一种Gram注意力模块，该模块将Gram矩阵作为物理先验引入以引导特征净化，使网络能够抑制与泄漏相关的分量，同时保留每个网格的支撑相关信息。净化后的特征被传递给解耦头以推断JAS域中的分层稀疏支撑。最后，在由检测到的支撑构建的低维字典上执行LS以恢复路径增益。仿真结果表明，所提出的框架在所有基线中实现了最高的支撑检测F1分数和最低的信道估计归一化均方误差（NMSE），而消融研究验证了将Gram域物理先验与数据驱动表示学习相结合的益处。
英文摘要
As next-generation wireless systems migrate to high-frequency bands, extremely large-scale multiple-input multiple-output (XL-MIMO) is indispensable for combating severe path loss. However, the received path energy along the extremely large array aperture (ELAA) may exhibit spatial non-stationarity, rendering the conventional discrete Fourier transform (DFT) codebook inadequate for channel estimation (CE) due to its full-array angular representation. To address this mismatch, we propose a novel joint angle-subarray (JAS) codebook, where each codeword corresponds to an angular direction and a specific subarray. Spatially non-stationary CE is thus transformed into structured sparse support detection in the JAS domain. However, reliable support recovery remains challenging for conventional model-based methods. We therefore develop a signal processing and deep learning co-design framework, where data-driven JAS parallel support detection is followed by least-squares (LS)-based channel reconstruction. The matched filter response of each JAS grid is embedded into a high-dimensional feature vector. We then propose a Gram-attention module that incorporates the Gram matrix as a physical prior to guide feature purification, enabling the network to suppress leakage-related components while preserving support-relevant information of each grid. The purified features are passed to decoupled heads to infer the hierarchical sparse supports in the JAS domain. Finally, LS is performed over the low-dimensional dictionary constructed from the detected supports to retrieve path gains. Simulation results demonstrate that the proposed framework achieves the highest support detection F1-score and the lowest CE normalized mean square error (NMSE) among all baselines, while ablation studies verify the benefits of integrating the Gram-domain physical prior with data-driven representation learning.
Comments13 pages, 10 figures
