Mehrdad Khani, Mohammad Alizadeh, Jakob Hoydis, Phil Fleming
TL;DR本文介绍了一种基于深度学习和迭代软阈值算法的 MMNet MIMO 检测方案,通过该方案,可以在同等或更低的计算复杂度下,实现对具有空间相关性的实际信道的在线训练,并在性能上远远优于现有的深度学习方法。
Abstract
Symbol detection for Massive Multiple-Input Multiple-Output (MIMO) is a
challenging problem for which traditional algorithms are either impractical or
suffer from performance limitations. Several recently proposed learning-based
approaches achieve promising results on simple channel models (e.g., i.i.d.
Gaussian). However, their performance degrades signific
本文提出了一种基于模型驱动的深度学习网络用于多输入多输出(MIMO)检测,通过深度学习技术优化网络的可训练参数以提高检测性能;由于网络可训练变量的数量等于层数,因此可以在很短的时间内轻松训练网络,并且网络可以处理时变信道,通过数值结果表明该方法可以显著提高 Rayleigh 和相关 MIMO 信道下迭代算法的性能。