Clustering Algorithm based on Azimuth in MmWave Massive MIMO-NOMA System

Abstract

In order to resolve the massive access in future generation, NOMA (non-orthogonal multiple access) is exploited in the Massive MIMO communication system at the mmWave (millimeter wave) frequency to gain a very high capacity and access a very large number of users. The user clustering is key issue, which give impact to the performance of the system. Due to the directional transmission at mmWave frequency and space multiplexing of Massive MIMO, we propose k-means clustering algorithm based on azimuth, which takes use of the user’s azimuth to group users with the similar azimuths into the same cluster, while users with distinct azimuths into the different clusters. It cannot only increase the access number from spatial aspect, and also reduce the inter-cluster interference. The simulation results show that the proposed algorithms can realize directional clustering, which well achieves the desired goal.