基于导频复用的大规模MIMO无线传输理论方法研究

发布时间：2018-07-03 10:56

本文选题：大规模MIMO + 导频复用　；参考：《东南大学》2016年博士论文

【摘要】：为了适应持续呈指数增长的无线业务需求,未来移动通信系统需要在无线物理层传输等多方面实现新的突破。其中,在基站侧配备大规模天线阵列的大规模多输入多输出(MIMO, multiple-input multiple-output)无线传输能够显著提升系统频谱效率和功率效率,被认为是未来移动通信系统最具潜力的研究方向。信道状态信息的获取是MIMO无线传输的基础,目前大规模MIMO无线传输中信道状态信息大多通过周期性发送正交导频来获取,相应的导频开销将随着天线数目线性增长。在中高速移动场景中,导频开销将消耗掉大量的时频资源,构成系统瓶颈。实际室外大规模MIMO无线信道通常呈现能量集中特性,利用这一信道特性,可以提升导频资源的复用能力,降低导频开销。鉴于此,本论文开展基于导频复用的大规模MIMO无线传输理论方法研究。全文主要工作与贡献如下：首先,针对平衰落信道,提出了基于角度域导频复用的大规模MIMO无线传输理论方法。从实际物理信道模型出发,研究了空间相关瑞利衰落大规模MIMO信道特性,证明了当基站侧天线数目趋于无穷大时,信道协方差矩阵的特征向量取决于基站侧阵列响应矢量,特征值取决于信道角度功率谱,从理论上揭示了信道空间相关性与信道角度功率谱之间的关系。基于这一信道模型,进而证明了当复用导频的不同用户其空间到达角的区间互相不重叠时,信道估计均方误差之和能够达到最小值,从理论上证明了在角度扩展受限的空间相关大规模MIMO信道上实施导频复用的可行性。考虑到导频复用可能会导致信道估计性能下降,进而研究了导频复用下的鲁棒上下行传输,推导得到了基于信号检测均方误差之和最小准则的鲁棒上行多用户检测器与鲁棒下行多用户预编码器的闭式表达,证明了两者之间的对偶性。在此基础上,研究了信道估计与信号检测均方误差之和最小准则下的导频调度方法。证明了当复用导频的不同用户其信道空间到达角区间互相不重叠时,两类均方误差性能均达到最优,并基于该最优条件提出了一种基于角度域信道统计特征的低复杂度导频调度算法。数值仿真结果表明,所提出的角度域导频复用相比于传统的正交导频能够显著提升系统净频谱效率。其次,针对采用正交频分复用(OFDM, orthogonal frequency division mul-tiplexing)调制的宽带大规模MIMO传输,提出了基于角度时延域导频复用的大规模MIMO-OFDM信道状态信息获取理论方法。从实际物理信道模型出发,推导得到了大规模MIMO-OFDM信道空间频率相关阵和信道角度时延功率谱之间的关系。从理论上证明,当基站侧天线数目趋于无穷大时,不同用户信道空间频率相关阵的特征向量趋近于相同,而特征值取决于各自的信道角度时延功率谱,从理论上揭示了大规模MIMO-OFDM信道在角度时延域的近似稀疏特性。基于这一信道模型,提出了频域相移可调导频,研究了基于频域相移可调导频的信道状态信息获取理论方法,并从理论上证明,当经过频域相移导频调度后的各用户等效信道在角度时延域互相不重叠时,相应的信道估计与信道预测均方误差之和均能达到最小值。依据该信道状态信息获取最优条件,进而提出了一种低复杂度的频域相移导频调度算法。上述结果在单个和多个连续OFDM符号的场景下均进行了论证。典型移动场景下的数值仿真结果表明,所提出的相移可调导频相比于传统的相移正交导频能够显著提升系统频谱效率。最后,针对毫米波／太赫兹大规模MIMO系统,提出了基于逐波束时频同步的波束分多址传输理论方法。从大规模MIMO波束域物理信道模型出发,证明了当基站侧与用户侧配置的天线数目均充分大时,波束域信道元素趋向于具有统计不相关特性且元素方差取决于信道角度功率谱。此外,波束域信道元素的包络趋向于不随时间和频率起伏。基于这一波束域信道特征,提出逐波束时频同步方法。从理论上证明了与传统的天线域同步相比,所提出的逐波束同步能够同时缩减宽带大规模MIMO信道的等效时延扩展和多普勒频率扩展,且缩减因子近似等于用户端配备的天线数目。进一步将所提出的逐波束同步方法应用于大规模MIMO波束分多址传输,研究了基站侧与用户端射频链路数目受限情形下的波束调度问题,并提出了一种基于波束域信道统计特征的低复杂度波束调度算法。典型移动场景下的数值仿真结果表明,在波束分多址传输框架下,所提出的逐波束同步方法相比传统的天线域同步方法能够显著降低波束域信道的多径效应和多普勒效应,提升其导频复用信道估计性能和无线传输性能,进而提升对毫米波／太赫兹信道下用户终端移动性的支持。
[Abstract]:In order to adapt to the continuous exponential growth of wireless service demand, future mobile communication systems need to achieve new breakthroughs in many aspects, such as wireless physical layer transmission. Among them, large-scale multi input and multiple output (MIMO, multiple-input multiple-output) wireless transmission with large antenna arrays on the base station side can significantly improve the spectrum efficiency of the system. Rate and power efficiency are considered as the most potential research direction of future mobile communication systems. The acquisition of channel state information is the basis of MIMO wireless transmission. At present, the channel state information in large-scale MIMO wireless transmission is mostly obtained through the periodic transmission of orthogonal pilot, and the corresponding pilot overhead will increase linearly with the number of antennas. In the medium and high speed mobile scene, the pilot cost will consume a large number of time and frequency resources and constitute the bottleneck of the system. The actual outdoor large-scale MIMO wireless channel usually presents the characteristics of the energy concentration. Using this channel characteristic, the multiplexing ability of the pilot resource can be improved and the pilot cost can be reduced. In this paper, a large-scale MI based on pilot multiplexing is carried out in this paper. The main work and contributions of the MO wireless transmission theory are as follows: first, a large scale MIMO wireless transmission theory based on angular domain pilot multiplexing is proposed for flat fading channels. Based on the actual physical channel model, the characteristics of the space related Rayleigh fading large conventional mode MIMO channel are studied, and the number of antenna number in the base station side is proved. When the mesh tends to infinity, the eigenvector of the channel covariance matrix depends on the response vector of the base station array, and the eigenvalue depends on the channel angle power spectrum. The relationship between the channel spatial correlation and the channel angle power spectrum is theoretically revealed. The sum of the mean square error of the channel estimation can be minimized when the range of the angle is not overlapped. It is theoretically proved that the pilot multiplexing is feasible on a space related large-scale MIMO channel with limited angle expansion. In the downlink transmission, the closed expression of robust uplink multiuser detector and robust downlink multiuser precoder based on the sum of mean square error of signal detection is derived, and the duality between the two is proved. On this basis, the pilot scheduler under the minimum sum of mean square error of channel estimation and signal detection is studied. The method. It is proved that the performance of the two class mean square error is optimal when the channel space of the different users with multiplex pilots does not overlap each other in the channel space. Based on the optimal condition, a low complexity pilot scheduling algorithm based on the statistical characteristics of the angle domain channel is proposed. The numerical simulation results show that the proposed angle domain pilot multiplexing phase is proposed. Compared with the traditional orthogonal pilot, the net spectrum efficiency of the system can be significantly improved. Secondly, a large-scale MIMO-OFDM channel state information acquisition method based on OFDM (orthogonal frequency division mul-tiplexing) modulation based on wide-band and large-scale MIMO transmission is proposed. Based on the physical channel model, the relationship between the spatial frequency correlation matrix of a large MIMO-OFDM channel and the time delay power spectrum of the channel angle is derived. It is proved theoretically that when the number of the base station side antennas tends to infinity, the eigenvectors of the spatial frequency correlation array of different user channels are similar, and the eigenvalues depend on their respective channels. The angle delay power spectrum reveals the approximate sparse characteristic of the large scale MIMO-OFDM channel in the angle delay domain. Based on this channel model, the frequency domain phase-shift tunable pilot is proposed. The theoretical method of channel state information acquisition based on the frequency domain phase-shift adjustable pilot is studied. It is proved theoretically in the frequency domain phase shift pilot modulation. The sum of the corresponding channel estimation and the mean square error of the channel prediction can be minimized when the angle delay domain of each user after the degree is not overlapped. The optimal condition is obtained according to the state information of the channel. Then a frequency domain phase shift pilot scheduling algorithm with low complexity is proposed. The results are in a single and multiple continuous OFDM characters. The numerical simulation results in a typical mobile scene show that the proposed phase-shift pilot can significantly improve the spectral efficiency of the system compared with the traditional phase shift orthogonal pilot. Finally, a beamsharing multi address transmission based on time frequency synchronization is proposed for the millimeter wave / terahertz large-scale MIMO system. From the large-scale MIMO beam domain physical channel model, it is proved that when the number of antennas at the base station side and the user side is all large, the beam channel elements tend to have statistical uncorrelation and the element variance depends on the channel angle power spectrum. In addition, the envelope of the channel elements of the wave beam domain is not dependent on the time. And frequency fluctuation. Based on the characteristics of this wave domain channel, a beamforming time frequency synchronization method is proposed. Theoretically, compared with the traditional antenna domain synchronization, the proposed beamforming can simultaneously reduce the equivalent delay expansion and Doppler frequency expansion of the wide-band large scale MIMO channel, and the reduction factor is approximately equal to the user end allocation. The number of antennas is further applied to the large scale MIMO beam division multiple access (mmdma) transmission. The beam scheduling problem is studied under the limited number of radio frequency links in the base station side and the user side, and a low complex beamforming algorithm based on the statistical characteristics of the beam domain channel is proposed. The numerical simulation results show that the proposed beambeam synchronization method can significantly reduce the multipath effect and the Doppler effect in the beam domain channel, and improve the performance of the pilot multiplexing channel estimation and the wireless transmission performance, and then improve the millimeter wave / terahertz channel. The support of the mobility of the lower end user terminal.
【学位授予单位】：东南大学
【学位级别】：博士
【学位授予年份】：2016
【分类号】：TN919.3

【相似文献】