[论文解读] Channel Estimation for Extremely Large-Scale MIMO: Far-Field or Near-Field?

Mingyao Cui, Linglong Dai|arXiv (Cornell University)|Aug 17, 2021

Advanced MIMO Systems Optimization参考文献 26被引用 11

一句话总结

本文提出了一种极坐标域信道估计框架，用于超大规模MIMO（XL-MIMO），该框架同时考虑了角度和距离信息，克服了传统远场平面波前假设的局限性。通过利用极坐标域中的稀疏性，提出了基于网格的（P-SOMP）和非网格的（P-SIGW）算法，在近场XL-MIMO系统中显著提高了估计精度并降低了 pilot 开销，优于传统的角度域方法。

ABSTRACT

Extremely large-scale multiple-input-multiple-output (XL-MIMO) with hybrid precoding is a promising technique to meet the high data rate requirements for future 6G communications. To realize efficient hybrid precoding, it is essential to obtain accurate channel state information. Existing channel estimation algorithms with low pilot overhead heavily rely on the channel sparsity in the angle domain, which is achieved by the classical far-field planar wavefront assumption. However, due to the non-negligible near-field spherical wavefront property in XL-MIMO systems, this channel sparsity in the angle domain is not available anymore, and thus existing far-field channel estimation schemes will suffer from severe performance loss. To address this problem, in this paper we study the near-field channel estimation by exploiting the polar-domain sparse representation of the near-field XL-MIMO channel. Specifically, unlike the classical angle-domain representation that only considers the angle information of the channel, we propose a new polar-domain representation, which simultaneously accounts for both the angle and distance information. In this way, the near-field channel also exhibits sparsity in the polar domain. By exploiting the channel sparsity in the polar domain, we propose the on-grid and off-grid near-field channel estimation schemes for XL-MIMO. Firstly, an on-grid polar-domain simultaneous orthogonal matching pursuit (P-SOMP) algorithm is proposed to efficiently estimate the near-field channel. Furthermore, to solve the resolution limitation of the on-grid P-SOMP algorithm, an off-grid polar-domain simultaneous iterative gridless weighted (P-SIGW) algorithm is proposed to improve the estimation accuracy, where the parameters of the near-field channel are directly estimated. Finally, numerical results are provided to verify the effectiveness of the proposed schemes.

研究动机与目标

• 解决现有基于远场的信道估计在XL-MIMO中因近场球面波前效应导致的性能下降问题。
• 利用一种新型极坐标域表示中的稀疏性，联合建模角度与距离以实现精确的信道估计。
• 开发适用于高频XL-MIMO系统中大规模阵列的低pilot开销估计算法。
• 通过引入更精确的极坐标域框架，克服基于网格和非网格的角度域方法的局限性。
• 通过在真实近场条件下进行仿真，验证所提方案的有效性。

提出的方法

• 提出一种极坐标域表示方法，利用出发/到达角度（AoD/AoA）和距离联合建模信道响应，捕捉近场传播的球面波前特性。
• 提出一种基于网格的极坐标域联合正交匹配追踪（P-SOMP）算法，用于在极坐标域的预定义网格上估计信道参数。
• 开发一种非网格的极坐标域联合迭代无网格加权（P-SIGW）算法，可直接估计连续的AoD、AoA和路径增益，无需受网格限制。
• 采用预白化处理以减轻混合波束成形架构在估计过程中引入的有色噪声影响。
• 采用基于梯度的优化框架，通过基于接收信号失配的代价函数，迭代优化角度和距离的估计值。
• 推导出代价函数相对于角度和距离参数的梯度解析表达式，以实现高效的迭代更新。

Figure 1: XL-MIMO system with hybrid precoding.

实验结果

研究问题

• RQ1 近场XL-MIMO信道在联合角度-距离（极坐标）域表示下是否表现出稀疏性？
• RQ2 由于球面波前效应，传统角度域压缩感知方法在近场XL-MIMO中的性能退化程度如何？
• RQ3 极坐标域表示是否能够在高精度下实现XL-MIMO系统中的低pilot开销信道估计？
• RQ4 基于网格和非网格的极坐标域估计算法在精度和对非网格偏差的鲁棒性方面有何差异？
• RQ5 混合波束成形引起的有色噪声对近场信道估计有何影响？如何有效缓解？

主要发现

• 近场XL-MIMO信道在极坐标域表现出稀疏性，支持基于压缩感知的高效估计。
• 所提出的P-SOMP算法相比传统远场方法，显著降低了pilot开销，同时实现了高精度信道估计。
• P-SIGW算法通过避免网格失配问题，能够处理非网格参数，从而在估计精度上优于P-SOMP。
• 仿真结果表明，所提极坐标域方法在近场场景中显著优于传统角度域方法，尤其在高信噪比（SNR）和大规模阵列条件下。
• 引入预白化处理可提升混合波束成形系统中对有色噪声的鲁棒性，从而整体提升估计性能。
• 理论分析表明，感知矩阵的列相干性有界，且依赖于散射体之间的距离差，验证了稀疏恢复的可行性。

Figure 2: Comparison of the angular-domain representations between the far-field and near-field channels. The BS antenna number is $256$ and the carrier is $100$ GHz. For far-field, the user is 100 meters from the BS, and for near-field, the user is 5 meters away from the BS while $L=3$ paths are ch