In this letter, we present a lower performance bound for discrete Fourier transform (DFT)-based beamspace channel estimation (CE) in 64 antennas Massive Multiple-Input Multiple-Output (MIMO) receiver. The beamspace CE is implemented after antennas digital signal transformation to a priori estimated DFT sub-space resulting in less complexity. To estimate the performance, we calculate noise power after the beamspace CE unit by decomposing an ideal channel response into separate taps and transforming them into the beam domain. Then the minimum mean squared error (MMSE) method is employed with taps to estimate the CE error power. An artificial CE is calculated as a sum of the ideal beamspace channel and additive Gaussian noise with the same error power. The artificial CE is utilized in the MIMO detector and decoder to achieve performance bound. Simulation results are presented for the non-line-of-sight model of the 5G QuaDRiGa 2.0 channel.

中文翻译：

---

大规模MIMO中波束空间信道估计的性能下限

在这封信中，我们提出了在64天线大规模多输入多输出（MIMO）接收机中基于离散傅里叶变换（DFT）的波束空间信道估计（CE）的性能下限。在将天线数字信号转换为先验估计的DFT子空间之后，实现波束空间CE，从而降低了复杂度。为了评估性能，我们通过将理想信道响应分解为单独的抽头并将其转换为波束域，来计算波束空间CE单元之后的噪声功率。然后，将最小均方误差（MMSE）方法与抽头一起使用，以估计CE误差功率。人工CE计算为理想波束空间通道和具有相同误差功率的加性高斯噪声的总和。在MIMO检测器和解码器中利用人工CE来达到性能极限。