Effective throughput maximization of beamspace MIMO-NOMA with finite blocklength

Abstract

Non-orthogonal multiple access (NOMA) has recently been integrated into beamspace multiple-input multiple-output (MIMO) for improved capacity and throughput. To apply the promising beamspace MIMO-NOMA in practical millimeter-wave applications, finite blocklength has to be considered. Therefore, in this article, we consider the effective throughput as the performance metric, which incorporates both the transmission rate and error performance in the finite blocklength regime. For the considered downlink beamspace MIMO-NOMA system, we derive the expression of system effective throughput with given blocklength and decoding error probability. To maximize the system effective throughput and simultaneously provide the quality-of-service (QoS) guarantee of data rate for each user, the transmit precoding and power allocation are optimized. We first provide an effective precoding design to mitigate the inter-beam interference. For power allocation, we apply monotonic optimization to obtain a globally optimal solution, and further develop a low-complexity algorithm based on the principles of convex-concave procedure (CCP). Simulation results show that the proposed schemes achieve higher spectrum and energy efficiency compared to several baseline schemes, including the traditional resource allocation algorithm based on the infinite blocklength assumption, and the existing beamspace MIMO.