An Approximate Evaluation of BER Performance for Downlink GSVD-NOMA with Joint Maximum-likelihood Detector

Abstract

Generalized Singular Value Decomposition (GSVD) is the enabling linear precoding scheme for multiple-input multiple-output (MIMO) non-orthogonal multiple access (NOMA) systems. In this paper, we extend research concerning downlink MIMO-NOMA systems with GSVD to cover bit error rate (BER) performance and to derive an approximate evaluation of the average BER performance. Specifically, we deploy, at the base station, the well-known technique of joint-modulation to generate NOMA symbols and joint maximum-likelihood (ML) to recover the transmitted data at end user locations. Consequently, the joint ML detector offers almost the same performance, in terms of average BER as ideal successive interference cancellation. Next, we also investigate BER performance of other precoding schemes, such as zero-forcing, block diagonalization, and simultaneous triangularization, comparing them with GSVD. Furthermore, BER performance is verified in different configurations in relation to the number of antennas. In cases where the number of transmit antennas is greater than twice the number of receive antennas, average BER performance is superior.

Keywords

• generalized singular value decomposition (GSVD)
• joint maximum-likelihood
• joint modulation
• MIMO
• nonorthogonal multiple access (NOMA)