| The sixth generation(6G)of mobile communications,which is the key component in intelligent manufacturing,medical care,transportation,and other intelligent infrustructures,must further advance the fundamental technical metrics to satisfy the demands of ubiquitous interconnection,all-area coverage,environmental technology,and low carbon emissions.Millimeter-wave(mm Wave)massive multiple-input-multiple-output(MIMO)communication,one of the supporting technologies to provide ultra-high data rate,is facing severe path loss and blockage issues in practical development.Nevertheless,the existing techniques by adjusting the transceiver design to account for wireless channel fading,have insurmountable technical bottlenecks due to their disability to actively reconfigure complicated wireless environments.The focus of research in next-generation wireless communication is on how to overcome the negative impact of fading of mm Wave wireless channels with low-cost,low-energy,and low-complexity technologies,and to achieve sustainable growth of future wireless communication capacity.Intelligent reflecting surface(IRS)is a new kind of metasurface with a large number of nearly passive reflecting elements.IRS,which enables intelligent reconfiguration of the wireless signal propagation environment,is one of the key candidate components for6 G.By properly adjusting the reflecting coefficient of each reflecting element,the equivalent channel can be improved to effectively increase the achievable rate of terminals and reduce the probability of link blockage.Due to the passive nature of the IRS,IRS-assisted mm Wave massive MIMO systems face many new challenges.In particular:(1)How to design a low-complexity joint beamforming algorithm to jointly design the active beamforming of base station(BS)and passive beamforming of IRS.Besides,the performance limits of IRS-assisted mm Wave MIMO systems also need to be explored?(2)How to design a low-cost and low-complexity channel state information(CSI)acquisition algorithm to support the joint beamforming design.This dissertation presents an in-depth study of these two challenges,including the following four research areas:(1)Joint beamforming design for multi-IRS assisted mm Wave multiple-input-single-output(MISO)systems?(2)Joint transceiver and passive beamforming design for IRS-assisted mm Wave MIMO systems?(3)Compressed Sensing(CS)-based full CSI acquisition for IRS-assisted mm Wave systems?(4)Sparse encoding/decoding and phaseless recovery-based fast beam alignment method for IRS-assisted mm Wave systems.First,this dissertation investigates the design of joint beamforming for single-stream transmission of multi-IRS-assisted mm Wave MISO systems.To reduce the computational complexity,the joint beam beamforming design problem is simplified by exploiting the line-of-sight(LOS)transmission of the mm Wave signal as well as the approximate orthogonality between the array response vectors.A closed-form optimal solution for the reflection coefficient is obtained for a single IRS-assisted mm Wave system,and a nearly-optimal analytical solution is obtained for multi-IRS-assisted mm Wave MISO systems.The proposed scheme significantly reduces the required computational resources by achieving a linear computational complexity.Both theoretical and simulation results show that the received signal power under the proposed joint beamforming design is proportional to the squared number of reflecting elements? the deployment of multiple IRSs can effectively increase the system throughput and reduce the link blockage probability.Second,this dissertation investigates the joint beamforming design for multi-stream transmission in IRS-assisted mm Wave massive MIMO systems.In conventional MIMO communications,the sparse scattering characteristics and LOS-dominated transmission of mmm Wave communications can hardly support high-quality multi-stream transmission.To overcome this challenge and reap the spatial multiplexing gain,this dissertation considers introducing a single IRS to assist mm Wave MIMO systems,where both BS and user equipment(UE)employ massive antennas to support multi-stream transmission and jointly design a hybrid precoding/combining for BS/UE and passive beamforming for IRS to improve the quality of multi-stream transmission.In this dissertation,the joint beamforming design problem in a single IRS-assisted mm Wave massive MIMO system is decoupled into two independent subproblems by exploiting the inherent sparse characteristics of mm Wave channels.For the passive beamforming design of IRS,a manifold optimization-based fast search algorithm with linear computational complexity is proposed.Simulation results reveal that the proposed scheme exhibits an average reduction of 99.3% in average runtime while experiencing only a 4% loss in achievable rate compared to the optimal algorithm.A more detailed examination of the results suggests that the proposed approach has the potential to enhance the channel condition number and support the multi-stream transmission of high quality.Third,this dissertation investigates the full CSI estimation problem in IRS-assisted mm Wave MISO and MIMO systems,where CSI is a prerequisite to fully exploit the potential of IRS.To tackle the difficulty in full CSI acquisition and huge training overhead,this dissertation points out that,from the perspective of joint beamforming,the full CSI of the IRS-associated cascaded channel is able to fully reap the joint beamforming gain.By exploiting the sparse characteristics of mm Wave channels,the cascaded BS-IRS-UE channel can be recast as a sparse representation.Consequently,the cascaded channel full CSI estimation problem can be formulated into a classical sparse signal recovery problem,which can be solved by a CS-based recovery algorithm.Both theoretical analysis and numerical results reveal that the proposed method can achieve performance close to the counterpart under perfect CSI with only 7% of the training overhead compared to the conventional algorithm.Finally,this dissertation investigates fast beam alignment in IRS-assisted mm Wave systems.Full CSI involves channel estimation algorithms that are too complex to be practical.In order to combat this challenge,by exploiting the sparse scattering characteristics of the mm Wave channel,this dissertation proposes a fast beam alignment method based on the sparse measurement and phaseless sparse signal recovery theory,which only needs to obtain the dominant path information of the channel and does not need to obtain the full CSI.By designing the sparse measurement matrix of BS and IRS,a multi-beam searching method is conducted to probe the cascaded channel,and then a low-complexity and intersection-based beam estimation method is proposed to achieve fast beam alignment.A robust scheme is proposed for non-line-of-sight(NLOS)scenarios.The success rate of the proposed beam is further analyzed for LOS and NLOS scenarios,which verifies that the proposed scheme can achieve excellent performance with only a small training overhead.Simulation results further show that the proposed method outperforms existing schemes in both LOS and NLOS scenarios while achieving similar performance to the exhaustive beam training method with 5% of the training overhead. |
PDF Full Text Request