The article considers robust joint AP clustering and beamforming design with imperfect CSI in cell-free systems. Specifically:
An optimization model is built to jointly optimize AP clustering and beamforming with imperfect CSI, aiming to maximize the worst-case sum rate and minimize the number of AP clustering under power constraint and sparsity constraint. The intractable semi-infinite constraints caused by imperfect CSI are transformed into more tractable forms.
The proposed RJAPCBN algorithm is designed to efficiently map CSI to beamforming. It includes an adaptive AP clustering module that adaptively sets an AP clustering threshold between each AP and each user, effectively reducing the issue of long-range APs consuming resources while contributing little. The adaptive AP clustering module uses a differentiable threshold function to enable joint optimization with the beamforming mapping.
Numerical results show that the proposed RJAPCBN achieves higher worst-case sum rate under smaller AP clustering, with much lower computational complexity compared to traditional and other deep learning algorithms.
翻譯成其他語言
從原文內容
arxiv.org
深入探究