Cell-free massive multiple-input multiple-output (CF mMIMO) systems are characterized by having many more access points (APs) than user equipments (UEs). A key challenge is to determine which APs should serve which UEs. Previous work has tackled this combinatorial problem heuristically. This paper proposes a sparse large-scale fading decoding (LSFD) design for CF mMIMO to jointly optimize the association and LSFD. We formulate a group sparsity problem and then solve it using a proximal algorithm with block-coordinate descent. Numerical results show that sparse LSFD achieves almost the same spectral efficiency as optimal LSFD, thus achieving a higher energy efficiency since the processing and signaling are reduced.
翻译:与用户设备相比,无细胞型大规模多投入多输出(CFMMIMO)系统的特征是有更多的接入点(APs)比用户设备(UEs)多。一个关键的挑战是如何确定哪些APs应该服务于UEs。先前的工作已经以超常方式处理了这一组合问题。本文建议为CFMIMO设计一个稀疏的大规模解码(LSFD)设计,以联合优化联系和LSFD。我们形成了一个群聚问题,然后用一个带有块状坐标底缘的准氧化算法来解决它。数字结果显示,分散的LSFD取得了与最佳LSFD几乎相同的光谱效率,从而在加工和信号减少后实现了更高的能源效率。