未来移动通信系统需持续提高频谱效率和能量效率，并提供全区域均匀覆盖。以用户为中心的去蜂窝（Cell-Free）大规模MIMO系统突破传统蜂窝系统设计理念，通过拉近用户与天线之间的距离，深度挖掘宏分集增益，大幅提升用户体验，是极具潜力的研究方向。然而，该技术正处在起步阶段，存在硬件性能受限、导频数量受限、计算能力受限和回程容量受限等多维资源受限问题，亟待深入研究。本项目利用信息论、凸优化、随机几何和随机矩阵等工具，针对硬件受限，提出基于信道空间强相关性的信息理论分析框架，为关键技术奠定基础；遵循移动通信系统传输流程，针对导频受限，提出基于图染色理论的导频优化策略；针对计算受限，提出基于深度学习的信道估计机制；针对容量受限，提出基于计算转发的回程链路设计方法，以期形成较为完整的去蜂窝大规模MIMO系统基础理论和关键技术，实现全区域均匀覆盖，以高质量用户体验满足未来移动通信系统持续增长的需求。

Future mobile communication systems should continue to improve the spectral and energy efficiency, and provide uniform service in a large area. In contrast to the conventional cellular system, cell-free massive multiple input and multiple output (MIMO) follows the user-centric idea to explore the macro diversity gain, which is obtained by shorting the distance between user equipment and antennas. However, the research of this system is just at early stage. The limited resource, such as hardware, pilot, computation, and capacity of backhaul, has not been addressed right now and need more research. We employ the information theory, convex optimization, random geometry, and random matrix theory to propose an information analytical framework of the cell-free massive MIMO system with hardware impairments over spatially correlated channels. Following the progress of mobile communication systems, we utilize the graph coloring theory to propose the pilot optimization method with limited number of pilots, use the deep learning theory to propose channel estimation scheme with limited computation ability, and utilize the compute and forward to design the backhaul with limited capacity. These results will contribute to the fundamental theory and key technologies of cell-free massive MIMO systems to provide uniform service in a large area, and fulfill the increasing need of future mobile communication systems with high-quality user experience.