基于大规模天线阵列混合波束成型的毫米波通信关键技术研究

The large-scale antenna array combined with beamforming can significantly increase the cellular coverage of millimeter-wave (mm-wave) communication systems. The large-scale antenna array, which is considered as the vital unit used to combat the severe path loss in mm-wave communications, hampers the implementation of full-digital beamformers because of the high hardware cost. As a result, the hybrid beamforming that accommodates digital beamformer with reduced dimension has been recognized as a promising technique to reduce the implementing cost. Unlike the full-digital beamformer, the analog units in the hybrid structure act as phase shifters with constant envelope. In other words, the analog units can not change the envelope but the phase of the input signal. Based on the hybrid beamforming structure, in this project we will study some key issues on the physical layer of mm-wave communication systems. Firstly, we will study the space sparsity of mm-wave propagation, and then set up the ray-tracing channel model. Especially, we will optimize the parameter configuration to incorporate different physical scenarios, such as indoor/outdoor, line-of-sight/non-line-of-sight, sunny/rainy cases. Based on the ray-tracing model, we will use two-dimensional beamspace MUSIC method to estimate the path directions and use the least square method to estimate the path gains. Unlike its element-space counterpart, the beamspace MUSIC can be disabled by inappropriate beamformers. We thus aim to derive the sufficient conditions to design the beamformers, which ensures the validity of the two-dimensional beamspace MUSIC. Secondly, combining the channel state information, we will propose new methods to design the hybrid beamformer with antenna sub-arrays. The sub-array structure further complicates the beamforming problem, because it accommodates the flexibility of using different types of sub-arrays. Thirdly, we will propose new methods to reduce the peak-to-average power ratio (PAPR) of the input signal, which helps to enhance the power efficiency of the power amplifiers. Especially, we will consider realistic constraints such as per-antenna peak power and signal cross-talks when designing the low-PAPR precoder. The results of this project will provide the theoretic foundation and technique support for mm-wave communication systems.

基于大规模天线阵列的波束成型技术由于能大幅提升毫米波微基站覆盖范围，已成为当前毫米波通信核心技术之一，且以模拟和数字器件相混合的波束成型结构因其电路工艺简化和制造成本低等方面的优势成为这一核心技术主要研究方向，但相应的信道估计、波束成型及预处理等物理层技术仍面临许多亟待攻克的关键难题。项目将基于该混合波束成型结构，利用毫米波信道的稀疏特性，基于波束域二维MUSIC的开环信道估计方法，研究波束成型设计的充分条件，以去除空间谱模糊效应并最大化可分辨路径数；结合信道状态信息，针对不同的模拟数字器件混合方案，以频谱效率为优化目标，提出低实现复杂度的混合波束成型新方法；将分别以误差函数和天线峰值功率为约束条件，研究混合波束成型结构中降低发射信号峰均比的预处理方法，以寻求提升发射端功率放大器能量转换效率的有效途径。项目研究将解决毫米波通信物理层面临的基础性问题，有效促进未来高频段移动通信技术发展。

基于大规模天线阵列的波束成型技术由于能大幅提升毫米波微基站覆盖范围，已成为当前毫米波通信核心技术之一，且以模拟和数字器件相混合的波束成型结构因其电路工艺简化和制造成本低等方面的优势成为这一核心技术主要研究方向，但相应的信道估计、波束成型及预处理等物理层技术仍面临许多亟待攻克的关键难题。项目将基于该混合波束成型结构，利用毫米波信道的稀疏特性，基于波束域二维MUSIC的开环信道估计方法，研究波束成型设计的充分条件，以去除空间谱模糊效应并最大化可分辨路径数；结合信道状态信息，针对不同的模拟数字器件混合方案，以频谱效率为优化目标，提出低实现复杂度的混合波束成型新方法；将分别以误差函数和天线峰值功率为约束条件，研究混合波束成型结构中降低发射信号峰均比的预处理方法，以寻求提升发射端功率放大器能量转换效率的有效途径。项目研究将解决毫米波通信物理层面临的基础性问题，有效促进未来高频段移动通信技术发展。