Deep learning-based user clustering in millimeter wave non-orthogonal multiple acccess communications

Abstract

A method and network node for deep learning based user clustering in millimeter wave (mmWave) non-orthogonal multiple access (NOMA) are disclosed. According to one aspect, a method includes determining a minimum number of clusters to be served by the network node subject to a first constraint that limits a total number of WDs in each cluster. The method also includes assigning WDs to each cluster of the minimum number of clusters while or after the minimum number of clusters is determined, the determined minimum number of clusters and WD assignments being for a first spatial distribution of the plurality of WDs.

Dependencies are inferred from the claim text.