The large bandwidth combined with ultra-massive multiple-input multiple-output (UM-MIMO) arrays enables terahertz (THz) systems to achieve terabits-per-second throughput. The THz systems are expected to operate in the near, intermediate, as well as the far-field. As such, channel estimation strategies suitable for the near, intermediate, or far-field have been introduced in the literature. In this work, we propose a cross-field, i.e., able to operate in near, intermediate, and far-field, compressive channel estimation strategy. For an array-of-subarrays (AoSA) architecture, the proposed method compares the received signals across the arrays to determine whether a near, intermediate, or far-field channel estimation approach will be appropriate. Subsequently, compressed estimation is performed in which the proximity of multiple subarrays (SAs) at the transmitter and receiver is exploited to reduce computational complexity and increase estimation accuracy. Numerical results show that the proposed method can enhance channel estimation accuracy and complexity at all distances of interest.

翻译：结合超大规模多输入多输出（UM-MIMO）阵列的大带宽使太赫兹（THz）系统能够实现每秒太比特的吞吐量。THz系统预期在近场、中场以及远场中运行。因此，文献中已引入了适用于近场、中场或远场的信道估计策略。本文提出了一种跨场（即能够在近场、中场和远场中运行）压缩信道估计策略。针对子阵列阵列（AoSA）架构，所提方法通过比较阵列间的接收信号，判断采用近场、中场还是远场信道估计方法更为合适。随后执行压缩估计，利用发射端和接收端多个子阵列（SA）的邻近性来降低计算复杂度并提高估计精度。数值结果表明，所提方法能够在所有感兴趣的距离范围内增强信道估计的精度和复杂度性能。