Universal Auto-encoder Framework for MIMO CSI Feedback

TL;DR

The proposed AE framework significantly reduces the HW complexity while providing comparable performance in terms of compression ratio-distortion trade-off compared to the naive and state-of-the-art approaches.

Abstract

Existing auto-encoder (AE)-based channel state information (CSI) frameworks have focused on a specific configuration of user equipment (UE) and base station (BS), and thus the input and output sizes of the AE are fixed. However, in the real-world scenario, the input and output sizes may vary depending on the number of antennas of the BS and UE and the allocated resource block in the frequency dimension. A naive approach to support the different input and output sizes is to use multiple AE models, which is impractical for the UE due to the limited HW resources. In this paper, we propose a universal AE framework that can support different input sizes and multiple compression ratios. The proposed AE framework significantly reduces the HW complexity while providing comparable performance in terms of compression ratio-distortion trade-off compared to the naive and state-of-the-art approaches.

Figures (11)

• Figure 1: Auto-encoder (AE) based framework for channel state information (CSI) feedback system.
• Figure 2: Comparison of objectives of existing works and ours. (a) Existing works aim to design multiple encoders ($\phi_1, \phi_2, \ldots$) providing a better trade-off between compensation ratio and distortion while the input size is fixed. (b) Our work aims to design a universal encoder ($\phi_{\mathtt{univ}}$) which can support various input and latent size to reduce the hardware (HW) complexity of the UE while it shows comparable performance in terms of compression-distortion trade-off.
• Figure 3: Three settings of partitioning $\mathbf{H}$ to reduce the input size of the AE and their NMSE performance.
• Figure 4: BLER performance of SVD-based beamforming with three types of CSI feedback.
• Figure 5: AE-based CSI feedback framework with the input space generalization when $K\in(2^{n-1},2^n]$ ($n$ is an integer).