Multi-antenna Transmission Techniques Research In Physical-Layer Multicast Systems

With the coming of the "Internet Plus" era, the types of wireless multimedia services are increasing rapidly and people have a higher requirement for wireless data transmission capacity. As a consequence, in future broadband wireless mobile communication system, the spectrum-efficient transmission technology will become a hot research topic. Based on the broadcast nature of radio links, physical-layer multicasting means that the transmitter sends the common message to all users simultaneously. In order to further improve the spectral efficiency of physical-layer multicast systems and to enhance the communication quality, the base station and users are equipped with multiple antennas. Under the multiple input multiple output (MIMO) configuration, this dissertation focuses on designing the transceiver with beamforing and precoding technology, which aims to improve the spectral efficiency and the quality of service of physical-layer multicast systems.Assume that full channel state information (CSI) of all users is available at the base station, two works have been done in physical-layer single-group multicast and physical-layer multi-group multicast scenarios:(1) For frequency-selective fading channels, the precoder for physical-layer single-group multicasting is investigated based on the receiver with finite impulse response, minimum mean-square-error and decision feedback equalization (FIR-MMSE-DFE), which aims to slove two problems, including maximizing the achievable rate of the system and optimizing the balance between data streams. For the first problem, an iterative FIR precoding algorithm is proposed. The convergence of the algorithm is verified and the computational complexity is analyzed. Simulations show that it has near-optimal achievable rate. To slove the second problem, the mean-square-error (MSE) matrix and signal to interference plus noise power ratio (SINR) of data streams are derived, and then an efficient unitary precoder is studied. Numerical simulations indicate that the proposed precoding design for single-group muticasting has better bit error rate (BER) performance than existing algorithms.(2) For frequency-selective fading channels, the beamforming algorithms for physical-layer multi-group systems are designed to slove the quality of service (QoS) problem and the max-min fair (MMF) problem. For solving the QoS problem, two low-complexity algorithms, i.e., the time-domain algorithm and the frequency-domain algorithm arc proposed. The relationship among key parameters, the computational complexity and the performance of the proposed algorithms are analyzed theoretically. To solve the MMF problem, the relationship between the QoS and MMF problems is discussed firstly, and then it is applied to design corresponding algorithms. Simulation results show that compared to the direct extension algorithm, the proposed time-domain algorithm and frequency-domain algorithm have little performance loss and huge computation reduction.As a natural extension of the above-mentioned investigations, what follows is physical-layer multi-group transmission technology under partial CSI feedback mode. With such a more practical assumption, this dissertation also carrys out the following research work:(3) For frequency-flat fading channels, a novel multi-group multicast beamforming scheme based on partial CSI feedback is presented. After defining the error angle between the accurate channel direction information (CDI) of each user and the chosen codebook, the lower bound of the average SINR of each user in the system is derived, and the QoS problem is formulated under this definition. Semi-definite relaxation (SDR) based and convex-optimization approximation based algorithms are proposed, and then compared in terms of computational complexity. To optimize the allocation of feedback resources, a step-by-step feedback scheme is explored, which is based on the fact that bottleneck users in multicast sytems have much greater influence on system performance than the rest of users. Numerical results demonstrate that proposed algorithms provide good performance with low complexity, and the step-by-step feedback method could optimize the allocation of feedback resources, thus improve system performance efficiently.