Flexible User Grouping for MIMO-NOMA Millimeter Wave Communication Systems

In this paper, we propose two novel user grouping strategies for the non-orthogonal multiple access (NOMA) technique based hybrid millimeter wave (mmWave) systems. In NOMA technology, user grouping is the first step to reduce the inter-group interference and will impact the design of hybrid beamforming. In contrast to some previous user grouping strategies that allocate equal or close number of users in each group, our proposed user grouping algorithms allow different number of users per group, which enables the users to get better beam gain. We propose two user grouping algorithms, the strong user dependent (SUD) user grouping algorithm and the Agglomerative Nesting (AGNES) clustering user grouping algorithm, in which a user is grouped to the beam with the largest beam gain. In the first algorithm, the transmitter selects the strong users as the beam centroids and allocates the rest weak users to the corresponding beam groups with the largest beam gain. The second algorithm directly forms groups based on the AGNES clustering algorithm without the predefined beam centroids. Simulation results demonstrate the effectiveness of our proposed user grouping strategies.