A Research On Beamforming Technologies In Terahertz Communications

This dissertation discusses the issue of improving the signal transmission efficiency and reliability of terahertz communication systems,which possess rich spectrum resources,high data rate,strong penetration,and strong directionality.Beamforming technology can effectively enhance signal transmission efficiency and reduce interference among communication users by generating strong directional beams through the use of an array antenna to achieve focusing effect,concentrating signal energy in the desired direction and thereby enhancing signal-to-noise ratio and interference resistance.Due to the high frequency and short wavelength of terahertz waves,terahertz array antennas can be designed very compactly,making beam control highly flexible.Thus,beamforming technology aimed at terahertz communication can effectively improve the transmission efficiency and reliability of terahertz signals,promote the development of terahertz communication systems,and play an important role in the future of wireless communication.Firstly,this dissertation summarizes the commonly used multiple input multiple output(MIMO)models in terahertz communication beamforming systems and proposes a more concise channel modeling method.For the terahertz channel,path gain is determined by both spatial free space attenuation loss and molecular absorption loss.This dissertation adopts the Friis formula and exponential attenuation model to model spatial free space attenuation loss and molecular absorption loss,respectively,and uses the high resolution transmission(HITRAN)database to determine the attenuation coefficient.Based on the proposed path gain model,this dissertation explores the relationship between transmission frequency and received power under the same size array,and introduces the differences between the channel under near-field conditions and far-field conditions,and how to judge whether it is in near-field/far-field state through Rayleigh distance.For antenna arrays with different geometrical structures,such as linear arrays,planar arrays,and multi-sided arrays,the corresponding array response vector expression is given.In addition,this dissertation outlines the radio frequency architecture of multi-antenna arrays,including fully digital architecture,fully analog architecture,and hybrid digital/analog architecture.Secondly,this dissertation elaborates on the concept of beam training,proposes a structural model for beam training: beam training consists of training protocol and codebook design.Particularly,this dissertation proposes a unilateral 3-tree search protocol,whose complexity is much lower than traditional protocols and requires no feedback information from the transmitter and receiver.Then,two high-quality codebook optimization algorithms are proposed: Successive Convex Approximation-Auxiliary Target Pursuit(SCA-ATP)and Sum of Symmetrical Array Response Vectors(S-SARV),respectively,for offline design and real-time design.Furthermore,based on the multi-sided array architecture,this dissertation develops a 3D beam training and tracking scheme that has application potential in dynamic beam management of terahertz communication.In terms of beam training,this dissertation proposes a new hierarchical codebook,where narrow beams ensure the highest worst-case performance,while wide beams have a smaller dead zone.Then,a low-complexity training protocol,called Grid-Based(GB)beam training,is proposed for finding the best narrow beam pairs.In terms of beam tracking,this dissertation proposes two tracking modes,collectively referred to as Grid-Based Hybrid(GBH)beam tracking,for achieving fast beam alignment of moving transceivers.Furthermore,the dissertation discusses the hardware feasibility of joint beamforming technology in terahertz intelligent reflecting surface(IRS)-assisted communication and introduces the applications of various emerging scenarios.For these scenarios,this dissertation proposes a beam training strategy under the terahertz IRS-assisted system and discusses how to use the angle information of the channel obtained through beam training to perform beamforming in a hybrid precoding architecture.In addition,the dissertation considers more generalized beamforming problems.For single-user scenarios,this dissertation proposes a Sum-Path-Gain Maximization(SPGM)criterion to simplify the problem-solving difficulty.Under this criterion,an Alternating Direction Method of Mul-tipliers(ADMM)approach is proposed to design the phase shift coefficient of the IRS.In multi-user scenarios,this dissertation proposes spatial orthogonality techniques,including the IRS-Complete Zero-Forcing(IRS-CZF)and IRS-Partial Zero-Forcing(IRS-PZF),to design the first part of the precoding matrix to eliminate interference between users.The second part of the precoding matrix is obtained by singular value decomposition(SVD)of the base station-user equivalent channel to maximize the achievable rate of the user.To solve the passive beamforming problem of the IRS,this dissertation proposes two efficient phase shift schemes: Water-Filling Segment Matching(WSM)and Phase Iterative Evolution(PIE),to balance the performance and complexity of phase shift design.Finally,this dissertation proposes a max-min beamformer for analog beamforming in terahertz wideband systems.This beamformer is designed to optimize analog beamforming to maximize wideband beam gain,i.e.,maximize the minimum beam gain on the signal frequency band.To solve this problem,this dissertation proposes an Augmented Lagrange Method(ALM)to optimize the simulated beamforming vector and proposes a Piecewise Response Vector(PRV)initialization method to obtain highquality convergence solutions.In addition,this dissertation investigates a more complex wideband codebook design problem to optimize beam zone divisions and beamforming vectors to maximize the wideband beam gain at the worst case.This dissertation proposes an optimal beam zone division method for wideband communication and expresses it in a semi-closed form.After determining the zone division,the ALM method to design beamforming vectors is adopted.When the system bandwidth is very large,severe beam split may occur at different frequency bands,and the max-min beamformer is no longer applicable.For this scenario,this dissertation derives the maximum effective bandwidth expression for beam splitting directions and no beam splitting effects and proposes a Beam Split Aggregation and Multiplexing(BSAM)concept for the Array of Subarray(Ao SA)architecture in wideband terahertz communication.