大规模MIMO系统传输技术和自适应调制研究

发布时间：2018-06-02 09:41

  本文选题：大规模MIMO + 分集传输　； 参考：《浙江大学》2017年硕士论文

【摘要】：随着4G进入规模商用的阶段,5G的研发也在如火如茶的展开。作为5G中的关键技术之一,大规模多输入多输出(MIMO)技术引起了广泛的关注。通过在基站端(BS)配备大规模天线阵列,大规模MIMO可以有效地提升系统的能量和频谱效率。随着基站天线个数的增加,小尺度衰落的影响被平均,不同用户到基站之间的信道渐进正交,基站端通过简单的线性处理技术就可以实现近似脏纸编码的性能。虽然具有很多潜在的优势,但大规模MIMO也依然面临诸多挑战,其中之一就是硬件的复杂性和巨大开销。如何利用大规模MIMO所带来的优势,简化传输技术和降低硬件的复杂度,是我们所研究的重点。首先,我们考虑了用户端配备两根天线的大规模MIMO系统的上行链路,研究了几种常见的分集传输方式的性能,如正交空时分组编码(OSTBC)、最大特征值一发射波束赋型(TB)和发射天线选择(TAS)。为了得到闭合的数学表达式,我们推导了可达和速率的下界,然后给出了当基站天线个数M趋于无穷时,近似的功率放大准则。解析表达式表明所研究的分集传输方式可以实现和单天线用户系统一样的功率准则,同时用户端的多个天线并不能带来明显的性能增益。因此,为了充分利用用户端多天线带来的空间维度,应当设计更加先进的传输方式。然后,考虑到大规模MIMO系统在实际实现过程中的硬件开销,通过能量有效的相移器将有限的射频(RF)链路和大量的基站天线连接在一起,即采用混合预编码的方案。针对目前多播系统中混合预编码研究的限制,我们提出了一种适用于任意信道(瑞利信道和毫米波信道等)的混合预编码方案。首先采用传统的半定松弛(SDR)方法求解,得到近似最佳的全数字预编码,然后通过最小化和全数字预编码的欧氏距离,基于交替优化技术,联合设计数字预编码和模拟预编码。仿真结果表明在有限个射频链路的条件下,我们所提出的混合预编码方案的性能非常接近全数字预编码的性能。最后,利用大规模MIMO系统带来的信道硬化(即小尺度衰落的影响被平均)的特征,我们提出了一种针对上行链路的多用户自适应调制方案,仅需要用户端缓慢变化的大尺度阴影衰落的信息,因此具有较低的复杂度。同时,我们给出了平均频谱效率(ASE)和误比特中断概率(BEO)的闭合表达式。和传统的快速自适应调制(FAM)相比,我们所提出的方案可以实现近似的性能,同时大大减小了反馈速率和开销。随后,为了进一步提升自适应调制系统的性能,提出了联合多用户自适应调制和功率控制的方案。
[Abstract]:As 4G enters the scale of commercial phase, the R & D of 5 G is also in the fire as tea. As one of the key technologies in 5 G, large-scale multi-input multi-output (Mimo) technology has attracted wide attention. Large scale MIMO can effectively improve the energy and spectral efficiency of the system by installing a large antenna array at the base station. With the increase of the number of base station antennas, the influence of small-scale fading is averaged. The channel between different users to the base station is gradually orthogonal, and the base station can achieve the performance of approximate dirty paper coding by simple linear processing technology. Although there are many potential advantages, large-scale MIMO still faces many challenges, one of which is the complexity and huge cost of hardware. How to make use of the advantages of large scale MIMO, simplify the transmission technology and reduce the complexity of hardware is the focus of our research. Firstly, we consider the uplink of a large scale MIMO system with two antennas on the client side, and study the performance of several common diversity transmission methods. For example, orthogonal space-time block coding (OSTBC), maximum eigenvalue with a transmit beam type (TB) and antenna selection (TAS). In order to obtain the closed mathematical expression, we derive the lower bound of the reachability sum rate, and then give the approximate power amplification criterion when the number of base station antennas M tends to infinity. The analytical expression shows that the proposed diversity transmission scheme can achieve the same power criterion as the single antenna user system, and that multiple antennas on the client side can not bring about significant performance gains. Therefore, in order to make full use of the spatial dimension brought by multi-antenna, more advanced transmission methods should be designed. Then, considering the hardware overhead of large-scale MIMO system, the limited RF RF link is connected with a large number of base station antennas through an energy-efficient phase shifter, that is, a hybrid precoding scheme is adopted. In view of the limitations of hybrid precoding in multicast systems, we propose a hybrid precoding scheme for arbitrary channels (Rayleigh channel, millimeter-wave channel, etc.). First, the traditional semi-definite relaxation SDR method is used to obtain the approximate optimal all-digital precoding, and then, by minimizing the Euclidean distance between all-digital precoding and all-digital precoding, combining digital precoding and analog precoding based on alternating optimization technique. Simulation results show that the performance of our hybrid precoding scheme is very close to that of full digital precoding under the condition of limited RF links. Finally, we propose a multi-user adaptive modulation scheme for uplink based on the characteristics of channel hardening (i.e. the effects of small-scale fading are averaged) brought by large-scale MIMO systems. It only needs the information of large-scale shadow fading which changes slowly at the user side, so it has low complexity. At the same time, we give the closed expressions of average spectral efficiency (ASE) and error bit interrupt probability (BER). Compared with the traditional fast adaptive modulation (Fam), the proposed scheme can achieve approximate performance and greatly reduce the feedback rate and overhead. Then, in order to further improve the performance of adaptive modulation system, a scheme of combining multi-user adaptive modulation and power control is proposed.
【学位授予单位】：浙江大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TN919.3
，

本文编号：1968300