| In order to meet the ever-increasing mobile data demands of users,wireless communications are moving towards "high frequency bands".Millimeter Wave(mmWave)Massive Multiple-Input Multiple-Output(Massive MIMO)has been considered as one of the key technologies in modern wireless communication systems,and has been used in the new generation cellular networks.By adopting a properly designed precoding/beamforming scheme,mmWave massive MIMO can generate huge array gains and support ultra-reliable and low-latency data transmission.However,when using traditional digital precoding/beamforming schemes,the base station needs to configure a dedicated RF link for each antenna,which results in extremely high hardware cost and power consumption.Therefore,how to ensure the superior performance of millimeter-wave massive MIMO while reducing the system’s demand for radio frequency links has become the focus and difficulty of current researches.Lens antenna arrays have energy-focusing properties that can convert a mmWave spatial channel into a sparser beam-domain channel.The so-called "sparseness" means that only a few elements in the channel matrix have large module values,and the modules of the remaining elements are approximately 0.Taking advantage of this physical property,a massive MIMO base station configured with a lens array can use a simple switching network to select these energy-focused beams from the beam-domain or beamspace channel for information transmission.This technique is called "beam selection".Beam selection can reduce the dimension of the effective channel matrix and the number of RF links required by the base station,which,thus,can reduce hardware cost and power consumption.At the same time,since the selected beam carries almost all the energy of the beam domain channel,beam selection will not cause a great loss to the system performance.Due to the above advantages,massive MIMO systems based on lens antenna arrays have received extensive attention from academia and industry.In this context,this paper focuses on the transmission design of millimeter-wave massive MIMO systems relying on lens arrays.More specifically,we consider three typical MIMO communication scenarios:narrowband channels,wideband channels,and wiretap channels,and discussed the problems of user scheduling,linear precoding,security features,etc.The specific research content and innovation points of this paper are as follows:1)The joint optimization problem of beam selection and user scheduling of narrowband downlink channel is studied,and two low-complexity and highefficiency joint optimization schemes based on bipartite graph model are proposed to optimize the downlink sum rate.Aiming at the problem that there are too many active users in the system,which leads to strong interference among users,an efficient joint design scheme of beam selection and user scheduling is proposed.With the goal of maximizing the system sum rate,this paper firstly analyzes the influence of the optimal number of served users on the system sum rate.Then,the bipartite graph model is used to model the joint design problem of beam selection and user scheduling,and a joint optimization scheme based on stable matching is proposed.In order to further reduce the computational complexity of the algorithm,this paper proposes a joint design scheme based on greedy algorithm.Simulation results show that:the proposed scheme can dynamically adjust the number of active users in the system,and can achieve the same sum rate performance as the exhaustive algorithm;compared with existing methods,the proposed stable matching method can improve the downlink sum rate by 50%and the computational complexity is only 6%of that of the exhaustive algorithm;compared with the existing methods,the proposed greedy algorithm can be improve the downlink sum rate by 40%and its computational complexity is only 2%of that of the exhaustive algorithm.2)The joint design of beam selection matrix on the base station side and radio frequency precoding on the user side of the multiple access channel is studied,and two joint optimization schemes based on the fractional programming technique are proposed to optimize the uplink system rate.Aiming at the problem of limited spatial freedom due to the configuration of a single antenna on the user side for transmission,a low-hardware-cost as well as low-energyconsumption transmission structure with a single radio frequency chain and multiple antennas on the user side is proposed.Under this configuration,an uplink sum rate maximization problem is formulated.The resultant problem has a complicated fractional structure,contains multiple nonconvex constraints,and is a nondeterministic polynomial-hard problem.In order to solve this problem,this paper adopts the framework of fractional programming capitalizing on the manifold optimization-based,stepwise refinement-based,and penalty-based methods,and proposes two effective joint design schemes for beam selection and radio frequency precoding.Simulation results show that compared with the existing schemes,the proposed joint design scheme based on stepwise refinement-based method can improve the system uplink sum rate by more than 30%in the high signal-to-noise ratio regime;in addition,compared with the joint design scheme based on the stepwise refinement algorithm,the joint design scheme based on penalty-based method can achieve a higher sum rate,but the stepwise refinement algorithm involves lower computational complexity.3)The joint design of beam selection and linear precoding for wideband channels is studied,and two joint optimization schemes based on signal-to-leakage-plusnoise ratio maximization criteria are proposed to optimize the downlink sum rate of wideband systems.Aiming at the problem that the system throughput is seriously influenced by the phenomenon of "beam squint" in the lens antenna array-aided wideband system,this paper adjusts the beam focusing direction by introducing a small number of phase shifters,and optimizes the baseband precoding,radio frequency precoding,and beam selection matrix to improve the sum rate of wideband systems.In order to simultaneously suppress the interference and enhance the signal,this paper designs the baseband precoding by maximizing the signal-to-leakage-plus-noise ratio,and obtains its closed-form solution through the generalized Rayleigh quotient theorem.Based on this,approximated expressions of the downlink sum rate and its lower bound are derived to simplify the analyses.Then,to achieve the two goals including "maximizing the approximation of the sum rate" and "maximizing the approximation of the sum rate lower bound",two joint design schemes are proposed under the framework of alternating optimization.More specifically,the radio frequency precoding is designed by the manifold optimization-based method,whereas the beam selection matrix is designed by the stepwise refinement algorithm.Simulation results show that compared with the existing scheme,the proposed two schemes can significantly improve the downlink sum rate of wideband systems.Moreover,in the low signal-to-noise ratio regime,the sum rate achieved by the sum rate lower bound-based design is higher than that achieved by the sum rate-based design,and vice versa in the high signal-to-noise ratio regime.In addition,the sum rate lower bound-based design involves lower computational complexity than the sum rate-based design.4)The joint design of beam selection and secure linear precoding for wiretap channels is studied,and several low-complexity schemes are proposed to enhance the secrecy capacity under the condition where the transmitter knows the eavesdropper’s statistical channel state information and instantaneous channel state information,respectively.Aiming at the problem of secure transmission in the wiretap channel relying on lens antenna array,this paper considers two typical scenarios:(1)the transmitter,Alice,knows the eavesdropper’s instantaneous channel information,(2)Alice knows the eavesdropper’s statistical channel information.For both cases,low-complexity algorithms are proposed to jointly design the beam selection matrix and the secure linear precoding in order to maximize the secrecy transmission rate subject to power constraints and minimize the transmission power under the guarantee of secrecy rate,respectively.Particularly,when Alice knows eavesdropper’s instantaneous channel state information,closed-form solutions of the optimal secure precoding vectors can be derived by using the generalized eigenvalue decomposition.Based on this fact,a twostage-based optimization framework is adopted to design the beam selection matrix and a stepwise refinement-based algorithm is proposed.As for the case where the transmitter knows the eavesdropper’s statistical channel state information,the optimal secure precoding vector can be only calculated numerically.On this condition,an alternating optimization-based method is proposed to design the secrecy precoding vector and the beam selection matrix iteratively.Simulation results show that compared with the existing schemes,the proposed joint design can significantly improve the secrecy performance of the wiretap channels for both the considered two cases. |
PDF Full Text Request