The Role of Deep Learning in NOMA for 5G and Beyond Communications

Abstract

In the coming future, it is obvious that the wireless networks will be congested with massive amounts of data traffic with the increasing number of users. Current multiple access techniques will certainly not have the capability to efficiently serve in the massively congested scenarios. In recent times, nonorthogonal multiple access (NOMA) has been recognized as an immensely potential technique for 5G and beyond communications that can increase spectral efficiency to a greater extent serving a vast number of users. However, several circumscriptions are observed in NOMA such as the requirement of a perfect channel state information in the transmitter and high computational complexity in the receiver. The use of deep learning (DL) techniques is a great resolution to deal with the challenges. This paper discusses the applications of the DL methods in NOMA for 5G and beyond communications. Firstly, the deep neural networks that are employed in NOMA are listed up and discussed. After that, their functions are studied specifically focusing on how to improve the NOMA performance. Finally, the possible future challenges and research issues are identified at the end of the paper.