Energy-Efficiency With Massive MIMO MU-NOMA in Symbiotic BackCom IoT Networks

In this letter, we develop a two-stage framework, based on both analytical modeling and machine learning (ML), for the analysis and optimization of a communication setup where the primary receivers (PRs) of a massive multiple-input multiple-output (mMIMO) multi-user non-orthogonal multiple access (NOMA) enabled primary network (PN) coexist symbiotically with a secondary network (SN) of backscatter-enabled tag transmitters (STs). The PN provides radio frequency signals to excite the semi-passive STs in their backscatter communication channel while gaining spatial diversity from the backscattering of the STs' desired signals. We aim to jointly optimize the primary transmitter (PT) beamforming, the PRs clustering, and the STs reflection coefficient to achieve maximal energy efficiency (EE). We propose an ML-based modified mean shift clustering for the PR clustering and an alternating optimization (AO) algorithm after the PR clustering to maximize the EE of Symbiotic radio network. We illustrate the proposed approach's superiority over conventional benchmarks with the help of simulation results.