基于压缩感知的大规模MIMO信道估计研究

Massive multiple-input multiple-out (MIMO) or large MIMO is a key essential technology for next-generation wireless communications, including mobile cellular system and millimeter wave system. To give full play to the great benefits of massive MIMO, channel estimation is dispensable. So far, channel estimation in massive MIMO is far from mature and therefore it needs further investigation. In view of this, we are motivated to investigate the issue of channel estimation in massive MIMO. In particular, we will focus on compressive sensing based channel estimation in massive MIMO systems. .The project contents include the following parts: .(1) In mobile cellular systems, the investigation of channel estimation exploiting the correlated sparsity patterns at base station side. .(2) In mobile cellular systems, the investigation of channel estimation exploiting the correlated sparsity patterns among multiple mobile users’ channels. .(3) In millimeter wave systems, the investigation of channel estimation considering hybrid analog/digital MIMO architectures and lens-based antenna array. .(4) The investigation of massive MIMO channel estimation taking into account non-ideal radio frequency (RF) circuits. The main RF non-idealities considered include in-phase quadrature (I/Q) imbalance and phase noise.

大规模MIMO技术是新一代无线通信系统中不可或缺的一项关键技术，面向的应用包括移动蜂窝系统和毫米波系统。为了充分发挥大规模MIMO的优势，信道估计必不可少。然而，目前大规模MIMO信道估计的研究还很不成熟，亟待深入。因此本项目将聚焦于基于压缩感知的大规模MIMO信道估计研究。.本项目研究内容包括：.(1) 移动蜂窝系统中，利用基站端信道稀疏特征相关的信道估计研究。.(2) 移动蜂窝系统中，利用多个移动用户信道间稀疏特征相关的信道估计研究。.(3) 毫米波系统中，针对模拟数字混合MIMO体制和透镜天线阵的信道估计研究。.(4) 存在射频非理想特性情况下，大规模MIMO信道估计研究。考虑的射频非理想特性主要包括I/Q通道不平衡和相位噪声。

大规模MIMO技术是新一代无线通信系统中不可或缺的一项关键技术，面向的应用包括移动蜂窝系统和毫米波系统。为了充分发挥大规模MIMO的优势，信道估计必不可少。本项目中，针对sub-6GHz频段（即移动蜂窝系统中常用的6GHz以下频段）和毫米波频段，基于压缩感知技术，展开大规模MIMO信道估计研究。研究内容和成果，聚焦在4个方面：.(1) Sub-6GHz频段大规模MIMO信道估计；.(2) 毫米波频段大规模MIMO信道估计；.(3) 采用透镜天线阵列的毫米波MIMO信道估计；.(4) 通信系统硬件非理想特性的估计、补偿/抑制。.研究得到的一系列算法，能够有效考虑sub-6GHz频段和毫米波频段信道的稀疏性、sub-6GHz信道在基站端和用户端的相关性、sub-6GHz信道在基站端的非平稳性、毫米波系统中的模拟/数字混合MIMO架构、采用透镜天线阵列的毫米波MIMO系统特点、和硬件非理想特性的影响，在导频资源开销、计算复杂度和性能上实现合理的均衡，明显优于文献中已有的同类型算法，甚至可以逼近最优的理论界。