Meta-Learning-Based Fronthaul Compression for Cloud Radio Access Networks

TL;DR

Simulation results show that the proposed two-stage neural network can perform close to the fully cooperative global CSI based benchmark with significantly reduced communication overhead for both the uplink and the downlink.

Abstract

This paper investigates the fronthaul compression problem in a user-centric cloud radio access network, in which single-antenna users are served by a central processor (CP) cooperatively via a cluster of remote radio heads (RRHs). To satisfy the fronthaul capacity constraint, this paper proposes a transform-compress-forward scheme, which consists of well-designed transformation matrices and uniform quantizers. The transformation matrices perform dimension reduction in the uplink and dimension expansion in the downlink. To reduce the communication overhead for designing the transformation matrices, this paper further proposes a deep learning framework to first learn a suboptimal transformation matrix at each RRH based on the local channel state information (CSI), and then to refine it iteratively. To facilitate the refinement process, we propose an efficient signaling scheme that only requires the transmission of low-dimensional effective CSI and its gradient between the CP and RRH, and further, a meta-learning based gated recurrent unit network to reduce the number of signaling transmission rounds. For the sum-rate maximization problem, simulation results show that the proposed two-stage neural network can perform close to the fully cooperative global CSI based benchmark with significantly reduced communication overhead for both the uplink and the downlink. Moreover, using the first stage alone can already outperform the existing local CSI based benchmark.

Figures (14)

• Figure 1: Uplink C-RAN system model. The transformation matrix $\mathbf{W}_b^{\mathrm{ul}}$ at RRH $b$ is designed using the proposed two-stage DNN to maximize the system utility. The quantizer $\mathcal{Q}_b^{\mathrm{ul}}\left(\cdot\right)$ is ignored when designing $\mathbf{W}_b^{\mathrm{ul}}$ and added after $\mathbf{W}_b^{\mathrm{ul}}$ is designed.
• Figure 2: Downlink C-RAN system model. The expansion matrix $\mathbf{W}_b^{\mathrm{dl}}$ at RRH $b$ is designed using the proposed two-stage DNN to maximize the system utility. The quantizer $\mathcal{Q}_b^{\mathrm{dl}}\left(\cdot\right)$ is ignored when designing $\mathbf{W}_b^{\mathrm{dl}}$ and added after $\mathbf{W}_b^{\mathrm{dl}}$ is designed.
• Figure 3: A block diagram of the proposed two-stage meta-learning algorithm for designing transformation matrices.
• Figure 4: Empirical CDF curve of uplink average per-user rate ($M=8$).
• Figure 5: Empirical CDF curve of uplink average per-user rate ($M=32$).