During the last decade, wireless communications has undergone a spectacular development with the introduction of multiple-input multiple-output (MIMO) technology. MIMO offers high data-rate and reliable transmission by exploiting the spatial multiplexing and diversity gains. Recently, massive MIMO has emerged as a new technology which scales conventional MIMO to a new level, offering unprecedented spectral efficiency and a significant reduction in transmit power. Massive MIMO is essentially a multiuser MIMO system where each base station (BS) employs a massive antenna array whose size is much greater than the number of users served. This setting is desirable as it allows extra degrees of freedom at the BS side, which can be used for interference cancellation. In massive MIMO, ea...[ Read more ]

During the last decade, wireless communications has undergone a spectacular development with the introduction of multiple-input multiple-output (MIMO) technology. MIMO offers high data-rate and reliable transmission by exploiting the spatial multiplexing and diversity gains. Recently, massive MIMO has emerged as a new technology which scales conventional MIMO to a new level, offering unprecedented spectral efficiency and a significant reduction in transmit power. Massive MIMO is essentially a multiuser MIMO system where each base station (BS) employs a massive antenna array whose size is much greater than the number of users served. This setting is desirable as it allows extra degrees of freedom at the BS side, which can be used for interference cancellation. In massive MIMO, each antenna element is connected to its own radio frequency (RF) chain, so the circuit power consumed by the overall RF chains scales linearly with the number of antennas. When the BS antenna size is scaled up, the increase in circuit power can become more prominent than the reduction in transmit power. As a result, the overall energy efficiency of massive MIMO systems can be poor. In this thesis, we aim to analyze and maximize the energy efficiency of massive MIMO systems.

In the first part, we will analyze the impact of pilot contamination on energy efficiency. Pilot contamination occurs when the number of users is greater than the pilot sequence length, which results in the corruption of the channel estimates at the BS during the uplink training phase. We shall investigate the energy efficiency performance in a single-cell massive MIMO system under both pilot contaminated and pilot uncontaminated cases. We will show that the energy efficiency in the pilot contaminated region deteriorates drastically with the number of antennas when compared to the case where there is no pilot contamination. Moreover, we will derive the closed form expression of energy efficiency as a function of the BS antennas and numerically determine the optimal number of antennas that can maximize the energy efficiency. Numerical results will then be verified by simulations.

In the second part of the thesis, we will aim to maximize the energy efficiency of massive MIMO using different power allocation schemes. Precoding, power allocation, user scheduling and antenna selection are some of the ways to improve energy efficiency. We will consider only power allocation for energy efficiency maximization. In particular, power allocation in single-cell massive MIMO under both pilot contaminated and pilot uncontaminated region will be studied. We propose three different power allocation algorithms; namely, heuristic, bisection and iterative. From the simulation results, we will show that the performance improvements which result from using the bisection and iterative algorithms are very close and better than those from the heuristic power allocation algorithm. However, the iterative algorithm is computationally more intensive than the bisection algorithm. Out of these three algorithms, the bisection algorithm was found to be most suitable for maximizing energy efficiency, both in terms of performance and complexity.