Adaptive Clustering of Users in Power Domain NOMA

By enabling multiple non-orthogonal transmissions, power domain non-orthogonal multiple access (PD-NOMA) potentially increases a system's spectral efficiency. This technique can become an alternative for future generations of wireless communication networks. The efficiency of this method fundamentally depends on two previous processing steps: an appropriate grouping of users (transmission candidates) as a function of the channel gains and the choice of power levels that will be used to transmit each signal. Thus far, the solutions presented in the literature to address the problems of user clustering and power allocation do not consider the dynamics of communication systems, i.e., the temporal variation in the number of users and the channel conditions. In order to consider these dynamic characteristics in the clustering of users in NOMA systems, this work proposes a new clustering technique based on a modification of the DenStream evolutionary algorithm, chosen for its evolutionary capacity, noise robustness and online processing. We evaluated the performance of the proposed clustering technique considering, for simplicity, the use of an already widely known power allocation strategy called improved fractional strategy power allocation (IFSPA). The results show that the proposed clustering technique can follow the system dynamics, clustering all users and favoring the uniformity of the transmission rate between the clusters. Compared to orthogonal multiple access (OMA) systems, the proposed model's gain was approximately 10%, obtained on a challenging communication scenario for NOMA systems since the channel model adopted does not favor a large difference in the channel gains between users.