Study of downlink multi-user mimo-noma using varied precoding, clustering, diversity and power allocation schemes with experimental validation

Abstract

Multiple-Input-Multiple-Output (MIMO) technology has gained significant attention in recent years to upgrade the performance of wireless communication systems. It allows the use of multiple antennas at both transmitting and receiving ends, thereby enabling more robust reception and increased system capacity. Non-Orthogonal Multiple Access (NOMA) is another promising radio access technology that outperforms the conventional multiple access techniques in order to maximize the spectral efficiency and minimize latency, thereby allowing users to share the same frequency, time, and code resources. A combination of MIMO and NOMA techniques, known as MIMO-NOMA, is said to be a propitious technology in further improving the system performance in various use cases, such as in beyond 5G broadband and machine-to-machine communication. In this thesis, a detailed performance analysis, based on MATLAB simulation, is presented for a multi-user MIMO power-domain NOMA (MU-MIMO-PD-NOMA) paradigm for varied combination of antenna diversity techniques, different types of precoding techniques and clustering methodologies using various modulation schemes and power allocation strategies. Also, a new precoding technique based on Singular Value Decomposition (SVD) precoding is proposed along with a novel clustering methodology and their performance is thereafter examined and compared with the existing precoding and clustering methods respectively. The modulation schemes used are namely, BPSK, QPSK and a combination of BPSK-QPSK, i.e., an Adaptive modulation technique. The power allocation strategies employed involve both fixed and dynamic power allocation. Under dynamic power allocation, Constrained Optimization and Convex Optimization algorithms are applied to the system respectively. Apart from simulation, a real-time test-bed implementation of (2x2) MIMO, without precoding and with different types of precoding, was performed on a Software Defined Radio (SDR) kit named WARP v3, in the 2.4GHz band. Physical implementation of two-user MIMO-PD-NOMA using fixed power allocation was also conducted on the WARP v3 board for the proposed and existing precoding schemes, at an experimentally determined optimum user position. The simulation and experimental performance parameters taken into consideration are Bit Error Rate, Sum Capacity, Spectral Efficiency and Constellation Diagram. The results were studied and a comparative analysis on grounds of the performance parameters, for different modulation schemes, is also presented in this thesis paper.