Massive Mimo Transmission With Statistical Channel State Information

With the rapid development of information society,there are tremendous changes and chal-lenges for mobile communication industry.Under the condition of limited spectrum resources,how to further improve the system transmission rate is an urgent issue to be addressed.Owing to large gains in power efficiency and spectral efficiency,massive multiple-input multiple-output(MIMO)is the key enabling technology for the 5th generation(5G)mobile communication sys-tem.However,in practical wireless communication system,the acquisition of accurate instan-taneous channel state information at transmitter(CSIT)is the bottleneck to achieve the potential benefits of massive MIMO.Moreover,with the popularization of mobile Internet and intelli-gent terminals,the application areas of wireless communication become increasingly diverse,and thus massive MIMO systems are expected to support data transmission services for different scenarios.Motivated by these reasons,this dissertation studies massive MIMO communication for several scenarios with only statistical CSIT.Focusing on the massive MIMO wireless communication systems with a passive evaes-dropper,the beam domain physical layer secure transmission is proposed.The optimality of the beam domain transmission is proved,and the corresponding power allocation optimiza-tion problem is investigated.Then,consider finite-alphabet input signals,the precoding design for multicell massive MIMO multicast transmission is investigated,and a lower bound on the achievable ergodic rate for finite-alphabet inputs is derived.Based on this lower bound,precod-ing design iterative algorithms for different system configurations are proposed.Moreover,in order to significant improve the spectral efficiency of underwater wireless communication sys-tems,the underwater acoustic transmission using large transceiver array apertures is proposed,which utilizes the high spatial resolution of massive MIMO to improve the system throughput.The major results and contributions of this dissertation are listed as follows.1.The beam domain secure transmission for massive MIMO communications is proposed.For massive MIMO systems with a passive eavesdropper,the optimality and power allo-cation algorithm of beam domain transmission is investigated.Focusing on the secure massive MIMO downlink transmission with only statistical channel state information(CSI)of legiti-mate users and the eavesdropper at base station,a lower bound on the achievable ergodic se-crecy sum-rate is introduced,from which the condition for eigenvectors of the optimal input covariance matrices is derived.The result shows that beam domain transmission can achieve optimal performance in terms of secrecy sum-rate lower bound maximization.For the case of single-antenna legitimate users,it is proven that allocating no power to the beams,where the beam gains of the eavesdropper are stronger,is optimal to maximize the secrecy sum-rate lower bound.This insight reveals the affect of the eavesdropper for beam domain power allocation,and it is beneficial to reduce computational complexity.Then,motivated by the concave-convex procedure(CCCP)and the large dimension random matrix theory,an efficient iterative and con-vergent algorithm is developed to optimize power allocation in the beam domain.Numerical simulations demonstrate the tightness of the secrecy sum-rate lower bound and the near-optimal performance of the proposed iterative algorithm.2.The multicellular massive MIMO multicast transmission with finite-alphabet inputs is proposed.For multicast transmission with finite-alphabet inputs,the precoding design in multicell massive MIMO systems is investigated.The users within each cell are interested in a common information and different cells provide distinct information.Focusing on the weighted max-min fairness(MMF)problem with only statistical CSI at the base station,the necessary conditions of the optimal precoding vectors are provided to maximize the minimum weighted achievable ergodic rate,and an iterative algorithm is proposed to optimize the precoding vectors.Then,to achieve lower computational complexity,a lower bound on the achievable ergodic rate is de-rived for finite-alphabet inputs.Considering the problem of the minimum weighted rate lower bound maximization,a CCCP-based algorithm is proposed,which is proved to converge to a local optimum.Furthermore,exploiting the channel characteristic in massive MIMO sys-tems,the optimal precoding vectors,which maximize the minimum weighted rate lower bound,are proved to be linear combinations of eigenvectors of transmit correlation matrices,and the original problem can be shifted into a lower dimensional space.Motivated by this insight,a relation-based algorithm is devised to obtain the optimal solution of the weighted MMF prob-lem by using the duality between the MMF problem and the quality of service(QOS)problem.Numerical results illustrate the tightness of the achievable ergodic rate lower bound and the significant performance of the devised algorithms.3.The underwater wireless transmission for massive MIMO communications is proposed.For shallow-water environment,the underwater acoustic massive MIMO system which deploys large array apertures at both the transmitter and receiver is investigated.A beam-based underwater acoustic massive MIMO channel model is proposed,while its asymptotic proper-ties are also analyzed.Based on this channel model,it is proved that the transmit design for rate maximization can be performed in a dimension-reduced space related to the channel taps.Then,with unlimited numbers of transducers,the beam-domain transmission is proved to be optimal for rate maximization.Furthermore,if the number of hydrophones also tends to in-finity,the optimal power allocation can be obtained just by the water-filling algorithm and the corresponding rate positively correlates with the number of channel taps for the high signal-to-noise-ratio(SNR)regime.Moreover,motivated by the random matrix theory,a low-complexity algorithm is devised to optimize the input covariance matrix for general cases.Simulation re-sults illustrate the high throughput achieved by massive MIMO and the effectiveness of the proposed algorithm.