| The emerging wireless applications and services introduce rapidly increasing requirements of high-datarate wireless access.The spectrum of low frequency(microwave,termed as μWave)band becomes strictly congested so that the μWave communication system cannot support the increasing high-data-rate broadband traffic.The millimeter wave(mm Wave)band has rich spectrum resources and can support broadband transmissions.Moreover,the small size of mm Wave antenna element enables the deployment of large scale array in compact physical size to provide high gain and directivity.However,due to the different characteristics of mm Wave signals and the implementation limitations of mm Wave components,techniques used in μWave communication system cannot be directly applied in mm Wave systems.This dissertation studies some key problems of the mm Wave communication systems,and the main work and contributions are listed as follows:Chapter 2 focuses on the derivation of coverage probability in mm Wave cellular networks.We propose an enhanced antenna model that expresses the mainlobe beamwidth and directivity gain as functions of the number of elements in the antenna array to model the directional antenna beamforming pattern.Based on the assumption that the beam alignment errors follow a truncated Gaussian distribution,the directivity gain with imperfect beam alignment is derived as a discrete random variable.We consider the random locations of user equipments(UEs)and base stations(BSs),and model the locations as homogenous Poisson point process(PPP).Then,the coverage probability is derived in a computationally tractable form.Numerical results validate the derivations,and illustrate the impact on the coverage performance of the BS intensity,the number of antenna elements and alignment errors.Chapter 3 analyzes the asymptotic achievable rate and spectral efficiency in the intelligent reflecting surface(IRS)based multiuser mm Wave networks.First,when the analog precoding codebooks used by the BS and IRS have infinite-resolution and full channel state information(CSI)of effective channels is fedback,we derive the asymptotic achievable rate and spectral efficiency.Then,the asymptotic achievable rate and spectral efficiency are derived when finite-resolution analog precoding codebook is used.We derive the upper bound of the spectral efficiency loss due to finite-resolution analog precoding and give an approximation in high signal-to-noise ratio(SNR)regimes.Furthermore,when the channel correlation based quantization(CQ)codebook is adopted to quantize the effective channels,the upper bound of the spectral efficiency due to limited CSI feedback and the approximation in high SNR regime are derived.The performance of CQ codebook is compared with the random vector quantization(RVQ)codebook.We also analyze the impact of the numbers of UEs and quantization bits.Numerical results validate the derivation and analysis.Chapter 4 focuses on the fast beam searching scheme in mm Wave communication for high-speed trains(HSTs).Leveraging the characteristics of mm Wave channels in HST scenarios,we model the sparse propagation paths as a birth and death process.Then,the sequential beam searching in channel estimation is formulated as a multi-armed bandit(MAB)problem in which the historical information is exploited to accelerate the estimation procedure.We use the transmission rate of each path as the reward and further propose an upper confidence bound(UCB)inspired beam searching scheme.The upper bound of the expected accumulated regret is derived to evaluate the scheme performance.Numerical results show that the proposed fast beam searching scheme can approach the theoretical limit in a few traverses with a few measurements.Chapter 5 jointly optimizes the user scheduling and transmit precoding problem in mm Wave and μWave dual-mode network.A concave expression is adopted to lower-bound the achievable rate of each UE with imperfect channel estimation.Based on the lower bound,an iterative algorithm is proposed to optimize the joint transmit precoding of mm Wave and μWave signals when all UEs’ schedule strategy is fixed.Then,this algorithm is extended to solve the cross-layer optimization problem in full dual transmission systems.For half dual transmission system in which all UEs can only operate in one frequency band,a branch-and-bound based algorithm and a rate based algorithm are proposed to optimize the joint user scheduling and transmit precoding.Numerical results validate the effectiveness and superiorities of the proposed algorithms. |
