Performance Analysis and Optimization in Massive MIMO Systems
No Thumbnail Available
Date
2021-03
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
Addis Ababa University
Abstract
Next generation networks are expected to support large volume of data traffic generated from emerging applications such as ultra-high speed video streaming, machine-tomachine (M2M) communication and the Internet of things (IoT). To handle this large volume of data traffic, these networks should employ technologies that utilize broad spectrum, offer higher cell density and high spectral efficiency. The spectral efficiency can be improved by increasing the transmitter power; introducing additional processing (such as deploying multiple antenna systems and advanced modulation techniques) in transceiver pairs that help to harvest energy; minimize multiuser-interference; and implementing innovative wireless planning and operation strategies that save energy. By deploying very large numbers of antennas at a base station (BS), which is called massive multiple input multiple output (MIMO), we can significantly improve the spectral efficiency of mobile networks. Besides, massive MIMO simplifies transmission processing, improves energy efficiency and reduces the required transmission power of the users. Due to those performance gains, massive MIMO has become an enabler for the deployment of 5G and beyond networks. In this PhD research, we study and analyze channel modeling, resource allocation and optimization techniques in massive MIMO systems. For this, first we analyze recent works on signal processing, channel modeling, channel estimation, resource allocation and optimization techniques in massive MIMO systems. In this regard, fundamentals of massive MIMO systems including channel capacity, spectral efficiency and energy efficiency have been studied. Closedform lower bound expressions are derived for the spectral efficiency and energy efficiency. Then, simulation results are provided to validate the theoretical analysis. Besides, performance analysis is done for linear detection and precoding techniques and then computationally efficient inverse approximation techniques are proposed for linear detection and precoding in massive MIMO systems. Specifically, Truncated Neumann series-based matrix inversion approximation techniques are formulated and probability of convergence, error of approximation and computationally complexity are analyzed. Then, we analyze achievable spectral efficiency of massive MIMO systems in realistic propagation environment under perfect and imperfect channel state information (CSI) scenarios. In particular, the effects of major large scale and small fading parameters including pathloss, shadowing, multipath fading, spatial channel correlation and impact of channel estimation have been investigated. Spectral efficiency analysis is done for uplink massive MIMO system under Rician fading channel model. Besides, by applying non-central to central Wishart approximation, closedform lower bound achievable rate expressions are formulated for massive MIMO systems in Rican fading channel model. Then, energy efficient power control and resource allocation algorithms have been proposed. For this, first by using large system analysis, analytical closedform lower bound expressions are derived for the achievable sum rate and appropriate power consumption model is formulated for the proposed massive MIMO systems. Then, by utilizing tools from fractional programming theory and sequential convex programming, energy efficient power control and resource allocation algorithms have been formulated. Further, the impacts of system and propagation parameters on energy efficiency have been evaluated. Particularly, the impacts of maximums transmitter power and minimum rate constraints of the users on global energy efficiency have been evaluated. The results show that the global energy efficiency increases with the maximum transmitter power constraint and decreases with the minimum data rate constraint. Finally, we analyze the performance of multicell massive MIMO systems in spatially correlated channel model. First, we study and evaluate fundamentals of multicell massive MIMO systems. In this regard channel modeling, power allocation and spatial resource allocation in multicell massive MIMO systems are considered. Besides, we evaluate the impacts of spatial correlation and pilot contamination. Important trade-offs and considerations on design and optimization of multicell massive MIMO systems has been studied. The impacts of system and propagation parameters are evaluated theoretically and via numerical simulation. The results show that spatial channel correlation has a major impact on channel hardening, favorable propagation, channel estimation quality and spectral efficiency of the system.
Description
Keywords
5G, massive MIMO, spectral effciency, energy effciency, optimization, resource allocation