Learning-Based Phase Compression and Quantization for Massive MIMO CSI Feedback with Magnitude-Aided Information

TL;DR

This paper introduces a learning-based framework that jointly optimizes phase and magnitude recovery for CSI feedback in Massive MIMO systems, leveraging magnitude-aided information to enhance accuracy in diverse environments.

Contribution

It proposes a novel end-to-end learning framework with a modified loss function for joint phase and magnitude CSI encoding, surpassing previous separate encoding methods.

Findings

Outperforms existing methods in phase recovery accuracy

Effective in both indoor and outdoor scenarios

Enhances overall CSI reconstruction quality

Abstract

Massive MIMO wireless FDD systems are often confronted by the challenge to efficiently obtain downlink channel state information (CSI). Previous works have demonstrated the potential in CSI encoding and recovery by take advantage of uplink/downlink reciprocity between their CSI magnitudes. However, such a framework separately encodes CSI phase and magnitude. To improve CSI encoding, we propose a learning-based framework based on limited CSI feedback and magnitude-aided information. Moving beyond previous works, our proposed framework with a modified loss function enables end-to-end learning to jointly optimize the CSI magnitude and phase recovery performance. Simulations show that the framework outperforms alternate approaches for phase recovery over overall CSI recovery in indoor and outdoor scenarios.