On The Performance Analysis And System Design For Large-Scale MIMO Systems

Large multiple-input-multiple-output (MIMO) are considered to be one of the key techniques in the next generation mobile communication systems to support high-speed mobile service. With large antenna array deployed in the base station (BS) or even in the users, the throughput can be substantially improved, and it can benefit from exploiting the statistical channel state-information (CSI). Moreover, besides the spec-tral efficiency, the energy efficiency can also be improved with large-scale MIMO deployment. However, the performance of the large MIMO systems is severely limited by the pilot resource. Based on the above motivation, we investigate the performance analysis and system design issues involving large-scale MIMO systems in this dissertation.Firstly, we consider uplink relay-assisted MIMO cellular systems with multiple relays deployed in each cell that perform the amplify-and-forward operations. We are interested in obtaining deterministic expressions for the ergodic sum rate of such systems. We first consider the large-scale MIMO dimension scenario, and obtain two asymptotic sum rate expressions, corresponding to the cases of fixed number of users and large number of users, respectively. We then consider the case of arbitrary MIMO dimension and large number of users, and obtain an upper and lower bound that the sum rate lies in between with high probability. The bounds are tight when the number of users is large. Both single-cell and multi-cell systems are considered, where the latter assumes full base station cooperation. Numerical experiments show that these deterministic sum rate expressions match well with the Monte Carlo simulation results. Therefore they can be useful tools for the design and analysis of relay-assisted MIMO cellular systems.Then, we investigate the sum-rate-optimal precoding for the uplink of the multi-cell large-scale MIMO systems. Specifically, we focus on transmitter precoder design based only on the statistical CSI. We first consider the partial cooperation system, where only the CSL is shared among cells, and its BS decodes its in-cell users by treating out-cell signals as colored Gaussian noise. We derive the ergodic sum rate and its deterministic approximation for the regime where both the number of total user antennas and the number of BS antennas are large. We obtain the necessary conditions to maximize the deterministic approximation of the sum rate under the transmit power constraint, based on which a gradient search algorithm is proposed to find the local optimal precoders. Further, a super cell system is also considered where users from all cells are jointly decoded. The deterministic approximation to the sum rate and optimal precoder design for such systems are given. Numerical experiments show that the proposed precoders achieve significant performance gains over systems with no precoding. Compared to the existing precoding methods that requires perfect CSI, the proposed precoders achieve similar sum rates and require only the statistical CSI.Next, we study the robust linear equalizer for the uplink of massive MIMO systems with multi-cell pilot reuse. We use the worst-case approach for robust design in order to combat pilot contamination. When the number of BS antennas is large, we build a novel model to indicate the relationship between the instantaneous channel matrix and the imperfect channel estimation in the presence of channel model uncertainty. Based on this model, we formulate the robust equalizer design problem into a min-max problem.-Further, we transform the min-max problem to an unconstrained one, and the optimality conditions are derived. With the resulting optimality conditions, two iterative algorithms and a simple approximate method are proposed to compute the optimal robust equalizers. Simulations are adopted to evaluate the performance of the proposed algorithms and approximate method. Compared to the conventional equalizers, the proposed robust equalizers achieve better bit-error-rate (BER) performance, especially in the regime of high signal-to-noise ratio (SNR), where pilot contamination is significant.Finally, we investigate the energy-efficient transmit scheme for the heterogenous network (HetNet) with large BS antenna array and soft-cell coordination. We consider the HetNet in the case of multiple macro BSs and small BSs coexisting with different transmit power. The HetNet uses soft-cell coordination strategy, i.e., only CSI is shared among the macro BSs and small BSs. We formulate the energy efficiency problem by minimizing the total transmit power with individual signal-to-interference-noise ratio (SINR) constraints for each user. The structure of the energy-efficient beamforming vectors is given according to the derived dual problem. We employ large-scale MIMO analysis to obtain the asymptotic dual problem so that the resulting internal parameters of the beamforming and transmit powers depend only on the statistic CSI to reduce the signal processing latency. As the asymptotic dual problem is not convex, iterative algorithms are proposed to compute the optimal beamforming parameters and transmit powers by converting the non-convex asymptotic dual problem into several convex sub-problems. Simulation results show the convergence of the proposed algorithms and the energy efficiency of the proposed transmit scheme.