LAUSR

Full-spreading non-orthogonal multiple access (FS-NOMA) is one category of the candidate technologies designed to support massive connectivity in wireless communication systems. Before it can handle the massive volume of user connections, it is important for the FS-NOMA to develop a receiver that successfully decodes target data from non-orthogonally overlapped receiving signals. However, the decoding performance of conventional interference-cancellation (IC)-based receivers is far from optimal because of error-propagation problems. To improve the decoding performance, we propose a novel FS-NOMA receiver based on the tabu-search (TS) algorithm which is a sort of machine-learning algorithm. Specifically, a novel TS mechanism and a diversification scheme are proposed to overcome the inherent adverse conditions of FS-NOMA systems which lead the TS algorithm to local optima. Simulation results demonstrate that the proposed TS-based receiver has decoding performance that is superior to that of the conventional IC-based receiver. The results also show that the proposed receiver accommodates a higher number of user connections with a given packet drop rate threshold.