Design Of Sum Rate Optimization Algorithm For Mmwave Full-Duplex MIMO Systems

Recently,with the rapid development of communication technologies and the continuous di-versity improvement of information services,the demand for communication is also increasing.The existing communication systems have been unable to meet the high demands for high through-put in the future.To address this problem,people put their emphasis on the millimeter wave in the high frequency band.In millimeter wave band,the available spectrum is much more than that band operating in forth-generation(4G)cellular systems or WiFi,resulting a higher potential transmis-sion capacity.However,some unfavorable qualities of millimeter wave communications such as large path loss and high system complexity limit its implementation.For the transmission rate problem in millimeter wave full-duplex multi-input multi-output(MIMO)communication system,this thesis propose effective optimization algorithms to improve the transmission rate of the system.This thesis first introduces the research background and the characteristics of the millimeter wave network,and then introduces some key technologies of the MIMO communication system,including the development and research status of beamforming and precoding techniques,full-duplex and relay techniques.Then we introduce the commonly used optimization theories and algorithms to provide theoretical basis for the subsequent algorithm design.Through joint beam-forming and precoding techniques,full-duplex and relay techniques,we have proposed a specific design scheme for millimeter wave full-duplex MIMO systems to increase the transmission rate of the system.In this thesis,a millimeter wave full-duplex relay system with channel state information(CSI)errors is considered.The hybrid beamforming matrices are jointly designed to maximize the total transmission rate of users.Explicitly,the transmit power constraints of the base station(BS)and relay station(RS),the residual self-interference power constraint of the RS,the per-user quality of service constraints and the unit-modulus constraints on the analog beamforming matrix elements are all taken into account.Since the resultant optimization problem is very challenging due to its highly nonlinear objective function and nonconvex coupling constraints,we first transform it into a more tractable augmented Lagrangian form.On this basis,we propose a joint optimization algorithm based on penalty dual decomposition(PDD).It is a double-loop algorithm,where the outer loop updates the penalty coefficient or the dual variable according to the constraint violation while the inner loop iteratively solves the augmented Lagrangian problem.The proposed algorithm converges to the set of stationary solutions of the original optimization problem.The algorithm is also extended to a more practical way with finite resolution phase shifters,where we design a codebook-based optimization algorithm.We show that our full-duplex hybrid millimeter wave MIMO relay system significantly outperforms both its nonrobust counterpart as well as the con-ventional half-duplex counterpart,and can effectively suppress the self-interference of full-duplex relay.A millimeter wave full-duplex multi-cell system is also studied,where we consider two dif-ferent structures:(1)fully-connected structure;(2)subarray structure.Aiming at maximizing the total system rate,a robust beamforming joint optimization method with CSI errors is proposed under the base station and uplink user energy constraints and the unit-modulus constraints on the analog beamforming matrix elements.The problem is nonconvex and very difficult to deal with.We first find a lower bound of the original problem,and approximate the rate optimization problem with CSI errors by maximizing the lower bound.Then we transform the lower bound maximiza-tion problem into an equivalent weighted minimum mean squared error problem and convert it into augmented Lagrangian form by introducing some auxiliary variables.We also adopt the PDD framework to design the algorithm,which converges to a set of stationary solutions.The simula-tion results show that our algorithm is robust of the CSI errors and achieve better rate performance.