The use of multiple-input multiple-output (MIMO) non-orthogonal multiple access (NOMA) based communication protocols is proposed and investigated for the uplink of wireless networks with buffered data-sources, which is the basis of the introduced medium access control (MAC)-layer protocol. To this end, the long-term average throughput is maximized by optimizing the set of users that transmit information at each time slot and their transmit power, the number of packets that are admitted in each user’s queue, and the transmission rates, assuming that the instantaneous channel state information is not available at the transmitters. Also, considering a receiver with multiple antennas, two detection techniques are used to mitigate the interference when two users are chosen to simultaneously transmit information in the same resource block, namely successive interference cancellation (SIC) and joint decoding (JD). More specifically, the outage probability for both considered techniques is derived in closed-from, which is a prerequisite for the derivation and the optimization of the throughput. The formulated multi-dimensional long-term stochastic optimization problem is solved by using the Lyapunov framework. Finally, simulation results verify the gains by using MIMO-NOMA as the basis of the next generation multiple access and illustrate the superiority of JD compared to SIC with respect to the number of the receiver’s antennas.

Cell-Free (CF) Massive MIMO (mMIMO) is a technology which can potentially augment not only the deployment of 5G, but also the deployment of "beyond 5G'' (B5G) wireless networks. The basic idea behind this “user-centric” arrangement is that many spatially dispersed access points (APs) coherently serve all nearby devices without being confined by cell boundaries. However, the cost for rolling out such systems, at a macro-level, may be significant. Radio stripes form a promising intermediate solution, which offers the potential of scalability at a reduced price. This paper investigates the uplink scenario of a CF massive MIMO (mMIMO) system, implemented with a limited-capacity radio stripe, which integrates a novel AP arrangement that fully exploits macro-diversity benefits. We also analyze a heuristic compare-and-forward (CnF) strategy, which, by comparing normalized minimum mean square error (N-MMSE) soft estimates, enables optimal dynamic cooperation clustering. Aiming at maximizing the spectral efficiency (SE) of a high percentage of users, we ensure that, under finite capacity constraints, our solution can guarantee better performance than existing radio stripe architectures when size of the system scales.