MIMO-NOMA系统中基于近邻传播的用户分簇算法

Affinity Propagation Based User Clustering Algorithm in MIMO-NOMA Systems

文中探讨了多输入多输出-非正交多址接入(Multiple-Input Multiple-Output Non-orthogonal Multiple Access,MIMO-NOM A)系统的用户分簇问题.针对现有用户分簇算法需要指定簇数的问题,提出了一种基于近邻传播的无监督机器学习用户分簇算法.仿真结果表明,提出的用户分簇算法在系统和速率上相较于对比算法具有显著的优势,同时算法不需要指定簇数,仅依赖于基站(Base Station,BS)处获取的信道状态信息(Channel State Information,CSI),便可将用户划分为多个簇,是一种方便且实用的算法.

A user clustering problem in multiple-input multiple-output non-orthogonal multiple access(MIMO-NOMA)systems is discussed.For the problem that the existing user clustering algorithms need to specify the number of clusters,this article proposes an unsupervised machine learning user clustering algorithm based on affinity propagation.Simulation results showed that the algorithm proposed in this paper has significant advantages over the comparison algorithm in terms of system sum rate.At the same time,the algorithm does not need to specify the number of clusters,and can only divide the user into multiple clusters by relying on the CSI obtained at the BS,which is a convenient and practical algorithm.