Learning Driven Resource Allocation and SIC Ordering in EH Relay Aided NB-IoT Networks
Learning Driven Resource Allocation and SIC Ordering in EH Relay Aided NB-IoT Networks
复制标题
EH 中继辅助 NB-IoT 网络中学习驱动的资源分配和 SIC 排序
DOI：
10.1109/lcomm.2021.3077635
复制
发表时间：
2021
期刊：
影响因子：
--
通讯作者：
L. Meng
中科院分区：
文献类型：
--
作者：
L. Qian;Chao Yang;Huimei Han;Yuan Wu;L. Meng
在窄带物联网（NB-IoT）网络中集成能量收集（EH）中继和非正交多址（NOMA）技术，可以有效提高网络的能量和频谱效率，改善边缘用户的服务质量。因此，我们在这封信中考虑EH中继辅助NOMA NB-IoT网络。为了减少NB-IoT设备之间的速率差异，我们的目标是通过联合优化通信资源分配和连续干扰消除（SIC）排序来最大限度地提高所有NB-IoT设备之间的数据速率比例公平性。考虑到该优化问题的非凸性，我们提出了一种基于深度强化学习的在线优化算法来获得次优解。仿真结果表明，与正交多址技术相比，该算法能有效提高NB-IoT设备间的比例公平性和总吞吐量。
Integrating the energy-harvesting (EH) relay and non-orthogonal multiple access (NOMA) technologies into narrow band internet of things (NB-IoT) networks can efficiently improve the energy and spectrum efficiency of the network and the quality-of-service of edge users. Therefore, we consider an EH relay aided NOMA NB-IoT network in this letter. To reduce the rate variance among NB-IoT devices, we aim to maximize the proportional fairness of data rate across all NB-IoT devices through jointly optimizing the communication resource allocation and successive interference cancellation (SIC) ordering subject to the minimum data rate requirements. Considering the non-convexity of this optimization problem, we propose a deep reinforcement learning based online optimization algorithm to obtain the sub-optimal solution. Simulation results demonstrate that the proposed algorithm can efficiently improve the proportional fairness and the total throughput among NB-IoT devices, in comparison with orthogonal multiple access techniques.