Decentralized Equalization With Feedforward Architectures for Massive MU-MIMO
Decentralized Equalization With Feedforward Architectures for Massive MU-MIMO
复制标题
DOI：
10.1109/tsp.2019.2928947
复制
发表时间：
2018-08
影响因子：
5.4
通讯作者：
Charles Jeon;Kaipeng Li;Joseph R. Cavallaro;Christoph Studer
中科院分区：
文献类型：
--
作者：
Charles Jeon;Kaipeng Li;Joseph R. Cavallaro;Christoph Studer
基于迫零 (ZF) 或线性最小均方误差 (L-MMSE) 均衡的线性数据检测算法可在大规模多用户多输入多输出 (MU-MIMO) 系统中实现接近最佳的频谱效率。然而，此类算法通常依赖于基站 (BS) 的集中处理，这会导致 1) 过多的互连和芯片输入/输出 (I/O) 数据速率以及 2) 高计算复杂性。分散式基带处理 (DBP) 将 BS 天线阵列划分为独立的集群，这些集群与单独的射频电路和计算结构相关联，以克服集中式处理的限制。在本文中，我们研究了具有前馈架构的去中心化均衡，该架构最大限度地减少了现有 DBP 解决方案的延迟瓶颈。我们提出了两种具有不同互连和 I/O 带宽要求的不同架构，它们融合了前馈网络中每个集群的本地均衡结果。对于这两种架构，我们考虑最大比率组合、ZF、L-MMSE 以及依赖于近似消息传递的非线性均衡算法。对于这些算法和架构，我们分析了相关的均衡后信噪比和干扰比。我们提供了多图形处理单元系统的参考实现结果，表明采用前馈架构的分散式均衡可实现 Gb/s 范围内的吞吐量，并且与集中式解决方案相比不会产生或仅产生很小的性能损失。
Linear data-detection algorithms that build on zero forcing (ZF) or linear minimum mean-square error (L-MMSE) equalization achieve near-optimal spectral efficiency in massive multi-user multiple-input multiple-output (MU-MIMO) systems. Such algorithms, however, typically rely on centralized processing at the base station (BS) which results in 1) excessive interconnect and chip input/output (I/O) data rates and 2) high computational complexity. Decentralized baseband processing (DBP) partitions the BS antenna array into independent clusters that are associated with separate radio-frequency circuits and computing fabrics in order to overcome the limitations of centralized processing. In this paper, we investigate decentralized equalization with feedforward architectures that minimize the latency bottlenecks of existing DBP solutions. We propose two distinct architectures with different interconnect and I/O bandwidth requirements that fuse the local equalization results of each cluster in a feedforward network. For both architectures, we consider maximum ratio combining, ZF, L-MMSE, and a nonlinear equalization algorithm that relies on approximate message passing. For these algorithms and architectures, we analyze the associated post-equalization signal-to-noise-and-interference-ratio. We provide reference implementation results on a multigraphics processing unit system which demonstrate that decentralized equalization with feedforward architectures enables throughputs in the Gb/s regime and incurs no or only a small performance loss compared to centralized solutions.