基于深度学习的信道估计

Channel estimation based on deep learning

目前深度学习方法在物理层通信中得到了广泛的应用,信道估计是物理层通信中的关键部分,与传统信道估计算法相比,深度学习方法在信道估计方面具有一定的优势。本研究介绍了深度学习中常见的神经网络模型,从模型驱动和数据驱动方式两个方面分别阐述了深度学习方法在信道估计中的应用。综述了将深度学习方法应用在信道估计中的最新研究进展。为了满足下一代移动通信的性能需求,讨论了智能物理层通信中信道估计面临的挑战与机遇。

Deep learning has been widely used in physical layer communication.Channel estimation is a key part of physical layer communication.Compared with traditional channel estimation algorithms,deep learning possesses some advantages in channel estimation.The common neural network models in deep learning was introduced,and the application of deep learning in channel estimation was explained from two aspects of model-driven and data-driven methods.The latest research progress in deep was learning applying to channel estimation has been reviewed.To meet the performance requirements of next-generation mobile communications,the challenges and opportunities of channel estimation in intelligent physical layer communications have been discussed.