| 基于元学习的短波MIMO信号检测 |
| |
| 引用本文： | 杨小石,卜方玲,王智勇,陈宁. 基于元学习的短波MIMO信号检测[J]. 计算机工程与设计, 2022, 43(1): 43-49. DOI: 10.16208/j.issn1000-7024.2022.01.006 |
| |
| 作者姓名： | 杨小石 卜方玲 王智勇 陈宁 |
| |
| 作者单位： | 武汉大学 电子信息学院,湖北 武汉 430072 |
| |
| 基金项目： | 国家重点研发计划项目课题基金项目(2018YFB2100503)。 |
| |
| 摘 要： | 针对短波场景下已有的MIMO检测算法性能不佳或复杂度太高的问题,提出一种基于元学习的MIMO检测方法MetaMMNet.将不同的信道矩阵下的MIMO检测看作为独立的任务,通过对大量不同任务的学习使得模型获得对于不同信道矩阵的泛化能力,能够快速适应变化的信道矩阵.在4 QAM、16 QAM和64 QAM调制方式下进行仿真...
|
| 关 键 词： | 短波通信 MIMO检测 迭代算法 神经网络 元学习 |
Meta learning based HF MIMO signal detection |
| |
| YANG Xiao-shi,PU Fang-ling,WANG Zhi-yong,CHEN Ning. Meta learning based HF MIMO signal detection[J]. Computer Engineering and Design, 2022, 43(1): 43-49. DOI: 10.16208/j.issn1000-7024.2022.01.006 |
| |
| Authors: | YANG Xiao-shi PU Fang-ling WANG Zhi-yong CHEN Ning |
| |
| Affiliation: | (School of Electronic Information,Wuhan University,Wuhan 430072,China) |
| |
| Abstract: | Aiming at the problems of poor performance and high complexity of existing MIMO detection algorithms in short-wave scenarios,a meta-learning based MIMO detection method MetaMMNet was proposed.MIMO detection under different channel matrices was regarded as an independent task,and the generalization ability of model for different channel matrices was obtained by learning a large number of different tasks.Simulation experiments were carried out under 4QAM,16QAM and 64QAM modulation.Experimental results show that MetaMMNet is better than MMNet(1 dB)and OAMPNet(2 dB)on short wave spatially-correlated channels.MetaMMNet can adapt to the variant channel more quickly than online learning MMNet,and has lower computational complexity compared with OAMPNet. |
| |
| Keywords: | HF communication MIMO detection iterative algorithm neural network meta learning |
| 本文献已被 维普 万方数据 等数据库收录！ |
|
