Deep Learning-Based Rate-Adaptive CSI Feedback for Wideband XL-MIMO Systems in the Near-Field Domain

TL;DR

This paper introduces WideNLNet-CA, a deep learning framework that adaptively compresses CSI feedback in wideband near-field XL-MIMO systems, addressing challenges of spherical wave effects and frequency dependence.

Contribution

It proposes a lightweight, rate-adaptive neural network architecture with a novel feature importance module for efficient CSI feedback in complex near-field wideband XL-MIMO scenarios.

Findings

Outperforms existing methods across various compression ratios and bandwidths.

Maintains fast inference and low storage requirements.

Effectively captures multi-scale channel features with reduced overhead.

Abstract

Accurate and efficient channel state information (CSI) feedback is crucial for unlocking the substantial spectral efficiency gains of extremely large-scale MIMO (XL-MIMO) systems in future 6G networks. However, the combination of near-field spherical wave propagation and frequency-dependent beam split effects in wideband scenarios poses significant challenges for CSI representation and compression. This paper proposes WideNLNet-CA, a rate-adaptive deep learning framework designed to enable efficient CSI feedback in wideband near-field XL-MIMO systems. WideNLNet-CA introduces a lightweight encoder-decoder architecture with multi-stage downsampling and upsampling, incorporating computationally efficient residual blocks to capture complex multi-scale channel features with reduced overhead. A novel compression ratio adaptive module with feature importance estimation is introduced to dynamically modulate feature selection based on target compression ratios, enabling flexible adaptation across a wide range of feedback rates using a single model. Evaluation results demonstrate that WideNLNet-CA consistently outperforms existing compressive sensing and deep learning-based works across various compression ratios and bandwidths, while maintaining fast inference and low model storage requirements.