ADAPTIVE PRECODING AND EQUALIZATION FOR MASSIVE MU-MIMO SYSTEM

The conventional multiple-input multiple-output (MIMO) system is incapable to meet the demand of very high data rate due to various up-coming applications such as machine-to-machine communications, video streaming, and internet of things (IoT) etc. To overcome these challenges, a promising candidate for the future generation is massive multiuser MIMO (MU-MIMO) due to its promise to provide both higher throughput and larger coverage. A massive MIMO system is the one that employs a very large antenna array at the base station which creates independent channels for each user and thus effectively increases system capacity. However, there are a number of challenges in the massive MIMO system such as pilot contamination due to pilot reuse within the cell, multi-user interference due to unwanted user signals, and inter symbol interference due to multipath channels etc. To combat all these issues, there is a need to design intelligent transmitter and receiver techniques for Massive MU-MIMO systems. In this context, adaptive precoder and equalizer play important role as they can improve the overall system performance. Typically, precoder and equalizer are designed either via supervised techniques by utilizing pilot sequences or via partial or perfect knowledge of the channel state information (CSI) at the base station and/or at the user’s mobile station (MS). The former technique suffers from pilot contamination while the later requires a lot of computation burden due to channel estimation process. More precisely, each user estimates the channel using pilot sequences transmitted from the base station, and then it feedbacks this CSI to the base station. This channel estimation overhead will be proportional to the number of base station antennas. Therefore, with massive MU-MIMO system, this is very inefficient. In this thesis, the design of blind algorithms for both the transmitter and receiver which requires only the knowledge of channel correlation matrix by using Indefinite Quadratic Forms is presented. For this, the first solution we provide is the design of an adaptive precoder with known equalizer. To do so, we first derive the outage probability in closed-form in terms of the eigenvalues of the channel correlation matrix. This is achieved by expressing the signal-to-interference-plus-noise-ratio (SINR) as an indefinite quadratic form and then by employing a recently developed approach which utilizes the Fourier transform expression for the unit step function. Next, we develop interior point based precoder by minimizing the derived outage probability expression. The second task of the thesis is to develop a method to provide simultaneous design of precoder and equalizer. Again, in order to achieve this, we first characterize the SINR of downlink multi-user-MIMO system in the presence of both multi-user interference and self interference and the closed-form expression for the outage probability is derived. This derived expression is then used to design an adaptive precoder and equalizer by employing an active-set based minimization technique. The third task of the thesis is to derive a closed-form expression for the probability of error for a BPSK modulated MU-MIMO system for the uplink scenario by employing again the methodology of indefinite quadratic form. Thus, an adaptive equalizer is designed that minimizes the probability of error expression via two approaches: interior point optimization and stochastic gradient optimization. The last task of the thesis is to provide coverage analysis of MU-MIMO system with spatial combining. Specifically, we consider the base station equipped with multiple antenna and users are equipped with a single-antenna mobile station. For the analysis purpose, we consider an uplink communication Scenario. Our analysis provides a thorough investigation of coverage probability with emphasis on the effect of users, the number of receive antenna elements and the noise power. Extensive simulations are presented to verify our theoretical findings.