This paper proposes a massive multiple-input multiple-output (MIMO) non-orthogonal multiple access (NOMA) system based on cognitive radio (CR) networks. To reduce the users interference, we propose an efficient user clustering algorithm. Within a cluster, we consider the user with the weak channel gain as a primary user (PU) and the user with the strong channel gain as a secondary user (SU). Given the quality of service for both PUs and SUs, we propose a power allocation (PA) method that maximizes the sum rate. The PA optimization problem is discussed into three scenarios which depend on the available transmit power. Thereafter, we derive a robust solution for the PA issue under imperfect channel state information (CSI) assumptions. Simulation results demonstrate that our proposed NOMA-CR scheme can significantly improve the spectral efficiency (SE). In addition, the proposed robust PA scheme for the imperfect CSI case reduces the performance loss due to the channel uncertainty.

中文翻译：

---

利用受 NOMA 启发的认知无线电进行大规模 MIMO 的用户集群和功率分配


本文提出了一种基于认知无线电（CR）网络的大规模多输入多输出（MIMO）非正交多址（NOMA）系统。为了减少用户干扰，我们提出了一种有效的用户聚类算法。在集群内，我们将具有弱信道增益的用户视为主要用户（PU），将具有强信道增益的用户视为次要用户（SU）。考虑到 PU 和 SU 的服务质量，我们提出了一种最大化总速率的功率分配（PA）方法。 PA 优化问题根据可用发射功率分为三种情况进行讨论。此后，我们在不完善的信道状态信息 (CSI) 假设下得出了 PA 问题的稳健解决方案。仿真结果表明，我们提出的 NOMA-CR 方案可以显着提高频谱效率（SE）。此外，针对不完美 CSI 情况提出的鲁棒 PA 方案减少了由于信道不确定性而导致的性能损失。