Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Feedback Design with VQ-VAE for Robust Precoding in Multi-User FDD Systems

Nurettin Turan, Michael Baur, Jianqing Li, Wolfgang Utschick

TL;DR

The paper tackles CSI feedback overhead in multi-user FDD massive MIMO by proposing a VQ-VAE-based feedback scheme that is tailored to the base-station environment. It uses an environment-aware VQ-VAE with a covariance-structured decoder output and a scalar embedding to generate a $B = N_\mathrm{L} \log_2 C$-bit feedback vector, and it draws samples from $\mathcal{N}_{\mathbb{C}}(\boldsymbol{\mu}_j, \mathbf{C}_j)$ within a SWMMSE loop to achieve robust multi-user precoding. Key contributions include the covariance-structured decoder, input pre-transform via $\mathbf{Q}\mathbf{P}^{\mathrm{H}}\mathbf{y}_j$, training over $0$–$20$ dB SNR, and demonstration on real-world data showing substantial gains over AE-based and DFT-codebook baselines with reduced feedback overhead. The approach promises practical impact by enabling high-performance precoding with fewer pilots and channel feedback, and it opens avenues for end-to-end optimization of both pilots and precoders. Overall, the work advances learned CSI feedback toward robust, environment-adaptive operation in 6G-era multi-user FDD systems.

Abstract

In this letter, we propose a vector quantized-variational autoencoder (VQ-VAE)-based feedback scheme for robust precoder design in multi-user frequency division duplex (FDD) systems. We demonstrate how the VQ-VAE can be tailored to specific propagation environments, focusing on systems with low pilot overhead, which is crucial in massive multiple-input multiple-output (MIMO). Extensive simulations with real-world measurement data show that our proposed feedback scheme outperforms state-of-the-art autoencoder (AE)-based compression schemes and conventional Discrete Fourier transform (DFT) codebook-based schemes. These improvements enable the deployment of systems with fewer feedback bits or pilots.

Feedback Design with VQ-VAE for Robust Precoding in Multi-User FDD Systems

TL;DR

The paper tackles CSI feedback overhead in multi-user FDD massive MIMO by proposing a VQ-VAE-based feedback scheme that is tailored to the base-station environment. It uses an environment-aware VQ-VAE with a covariance-structured decoder output and a scalar embedding to generate a -bit feedback vector, and it draws samples from within a SWMMSE loop to achieve robust multi-user precoding. Key contributions include the covariance-structured decoder, input pre-transform via , training over – dB SNR, and demonstration on real-world data showing substantial gains over AE-based and DFT-codebook baselines with reduced feedback overhead. The approach promises practical impact by enabling high-performance precoding with fewer pilots and channel feedback, and it opens avenues for end-to-end optimization of both pilots and precoders. Overall, the work advances learned CSI feedback toward robust, environment-adaptive operation in 6G-era multi-user FDD systems.

Abstract

In this letter, we propose a vector quantized-variational autoencoder (VQ-VAE)-based feedback scheme for robust precoder design in multi-user frequency division duplex (FDD) systems. We demonstrate how the VQ-VAE can be tailored to specific propagation environments, focusing on systems with low pilot overhead, which is crucial in massive multiple-input multiple-output (MIMO). Extensive simulations with real-world measurement data show that our proposed feedback scheme outperforms state-of-the-art autoencoder (AE)-based compression schemes and conventional Discrete Fourier transform (DFT) codebook-based schemes. These improvements enable the deployment of systems with fewer feedback bits or pilots.
Paper Structure (11 sections, 10 equations, 6 figures, 1 algorithm)

This paper contains 11 sections, 10 equations, 6 figures, 1 algorithm.

Table of Contents

  1. Introduction
  2. System Model and Channel Data
  3. Data Transmission Phase
  4. Pilot Transmission Phase
  5. Real-World Channel Data
  6. Proposed VQ-VAE-based Feedback Scheme for Robust Precoder Design
  7. Baseline Feedback Schemes for Precoding
  8. State-of-the-art AE and VQ-VAE-based Feedback Schemes
  9. Conventional DFT Codebook-based Feedback Scheme
  10. Simulation Results
  11. Conclusion

Figures (6)

  • Figure 1: Measurement campus in Stuttgart, Germany HeDeWeKoUt19.
  • Figure 2: Structure of the proposed VQ-VAE for robust precoder design.
  • Figure 3: The sum-rate over the SNR for a system with $B=40$ feedback bits, $J=8$MT, and $n_{\mathrm{p}}=8$ pilots.
  • Figure 4: The sum-rate over the number of pilots $n_\mathrm{p}$ for a system with $J=8$MT, $B=40$ feedback bits, and $\text{SNR}=15dB$.
  • Figure 5: The sum-rate over the number of MT $J$ for a system with $B=40$ feedback bits, $n_\mathrm{p}=8$ pilots, and $\text{SNR}=15dB$.
  • ...and 1 more figures