AI 中文总结
该研究提出特定站点多用户 MIMO 波束成形框架，利用特定站点信息，学习波束域观测到用户发射子空间的映射，在 CSI-RS 获取前推断用户可分性，构建紧凑组级 CSI 获取子空间，实现低开销、高有效速率及低用户侧处理负担。
AI 中文摘要
提出了一种低开销的特定站点多用户多输入多输出（MU-MIMO）波束成形框架。传统的有限反馈 MU-MIMO 在分组和波束成形前依赖信道状态信息参考信号（CSI-RS）传输和用户反馈，当天线维度和候选用户池很大时，这需要大量在线开销。为减轻此负担，所提框架利用捕获本地无线电传播特征的特定站点信息（SSI）。通过学习从低开销波束域观测到用户有效发射空间子空间的映射，基站可在高分辨率 CSI 获取前推断用户间可分性，并为选定用户构建紧凑的组级 CSI 获取子空间。这种特定站点设计可在标准有限反馈过程中实现，使用基于同步信号块（SSB）的参考信号接收功率（RSRP）指纹进行子空间推断，并使用 CSI-RS 反馈进行低维 CSI 细化。大量数值结果表明，所提框架可在 CSI-RS 获取前识别兼容用户组，在紧凑组子空间中保留大多数调度用户信道能量，并以显著更低的开销和用户侧处理负担实现比传统系统更高的有效速率。
英文摘要
A low-overhead site-specific multi-user multiple-input multiple-output (MU-MIMO) beamforming framework is proposed. Conventional limited-feedback MU-MIMO relies on channel state information reference signal (CSI-RS) transmission and user feedback before grouping and beamforming, which requires substantial online overhead when the antenna dimension and candidate-user pool are large. To reduce this burden, the proposed framework exploits site-specific information (SSI), which captures local radio propagation features. By learning the mapping from low-overhead beam-domain observations to effective transmit spatial subspaces of users, the BS can infer inter-user separability before high-resolution CSI acquisition and construct a compact group-level CSI acquisition subspace for the selected users. This site-specific design can be implemented within the standard limited-feedback procedure using synchronization signal block (SSB)-based reference signal received power (RSRP) fingerprints for subspace inference and CSI-RS feedback for low-dimensional CSI refinement. Extensive numerical results demonstrate that the proposed framework can identify compatible user groups before CSI-RS acquisition, preserve most scheduled-user channel energy in a compact group subspace, and achieve higher effective rates than conventional systems with significantly lower overhead and user-side processing burden.
