无蜂窝大规模MIMO中的大规模随机接入

摘要：研究了以用户为中心的无蜂窝大规模多输入多输出（MIMO）架构下的大规模随机接入方案。为实现可扩展架构，讨论了接入点（AP）与用户设备（UE）的关联以及AP的分簇方法。针对活跃用户检测（AUD），一种基于最大似然检测（ML）的方案被提出以获取活跃用户。通过调整阈值，可以得到精度不同的检测结果。利用AUD的检测用户集合，系统利用基于狄利克雷过程的稀疏贝叶斯学习（DP-SBL）完成信道估计（CE），该算法可以充分的利用AP的空间聚散特性，提高准确性。基于以上工作，我们提出了联合AUD和CE算法。仿真结果验证了所提方案在性能上的优越性。

关键词：大规模随机接入；无蜂窝大规模MIMO；活跃用户检测；信道估计

Abstract: A user-centric massive random access scheme under the context of cell-free massive multiple-input multiple-output (MIMO) architecture is investigated. To achieve a scalable architecture, the association between access points (APs) and user equipment (UE) as well as the clustering method for APs is discussed. Regarding active UE detection (AUD), a class of maximum likelihood (ML)-based schemes is proposed to obtain active UE set. By adjusting the threshold, detection results with varying accuracies can be achieved. Leveraging the detected user set from AUD, the system employs sparse Bayesian learning based on Dirichlet process (DP-SBL) for channel estimation (CE), effectively utilizing the spatial clustering characteristics of APs to enhance accuracy. Building on this, we propose a joint AUD and CE algorithm. Simulation results validate the superiority of the proposed approach in terms of performance.

Keywords: massive random access; cell-free massive MIMO; active UE detection; channel estimation.