Deep learning based channel state information (CSI) feedback in frequency division duplex systems has drawn widespread attention in both academia and industry. In this paper, we focus on integrating the Type-II codebook in the wireless communication standards with deep learning to enhance the performance of CSI feedback. In contrast to the existing deep learning based studies on the Release 16 Type-II codebook, the Type-II codebook in Release 17 (R17) exploits the angular-delay-domain partial reciprocity between uplink and downlink channels to select part of angular-delay-domain ports for measuring and feeding back the downlink CSI, where the performance of deep learning based conventional methods is limited due to the deficiency of sparse structures. To address this issue, we propose two new perspectives of adopting deep learning to improve the R17 Type-II codebook. Firstly, considering the low signal-to-noise ratio of uplink channels, deep learning is utilized to accurately select the dominant angular-delay-domain ports, where the focal loss is harnessed to solve the class imbalance problem. Secondly, we propose to adopt deep learning to reconstruct the downlink CSI based on the feedback of the R17 Type-II codebook at the base station, where the information of sparse structures can be effectively leveraged. Furthermore, a weighted shortcut module is designed to facilitate the accurate reconstruction, and a two-stage loss function that combines the mean squared error and sum rate is proposed for adapting to practical multi-user scenarios. Simulation results demonstrate that our proposed deep learning based port selection and CSI reconstruction methods can improve the sum rate performance compared with the traditional R17 Type-II codebook and deep learning benchmarks.

翻译：在频分双工系统中，基于深度学习的信道状态信息反馈已引起学术界和工业界的广泛关注。本文聚焦于将无线通信标准中的Type-II码本与深度学习相结合，以提升CSI反馈性能。与现有基于Release 16 Type-II码本的深度学习研究不同，Release 17（R17）中的Type-II码本利用上下行链路在角度-时延域的部分互易性，选择部分角度-时延域端口进行下行CSI测量和反馈，其中基于深度学习的传统方法因稀疏结构不足而性能受限。为解决该问题，我们提出两种采用深度学习改进R17 Type-II码本的新视角。首先，针对上行链路低信噪比问题，利用深度学习精确选择主导角度-时延域端口，并采用焦点损失函数解决类别不平衡问题。其次，提出在基站侧基于R17 Type-II码本反馈信息，采用深度学习重构下行CSI，从而有效利用稀疏结构信息。此外，设计加权捷径模块促进精确重构，并提出结合均方误差和和速率的双阶段损失函数以适应实际多用户场景。仿真结果表明，与传统R17 Type-II码本及深度学习基准方法相比，我们提出的基于深度学习的端口选择和CSI重构方法可提升和速率性能。