A Survey of power allocation techniques in NOMA: Research challenges and Future directions

Abstract

Non-orthogonal multiple access, often known as NOMA, is one of the viable ways to big capacity radio access. It provides a number of desired features, including better spectrum efficiency, making it an appealing choice. This piece places a focus on power-domain NOMA, in which successive interference cancellation (SIC) and superposition coding (SC) are the most essential functions at the transmitter and receiver, respectively. Following an analysis of many standard power allocation methods and the restrictions they impose, the authors of this article go on to describe a variety of innovative power distribution techniques that are based on machine learning. Approaches that are based on machine learning and deep learning produced performance that was considerably near to the optimal in terms of total capacity, although having significantly lower computing costs. Optimal performance would be attained by having the most overall capacity. Discussion of a number of potential future research avenues based on the use of deep learning in NOMA systems is the last step of the process.