Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Dual-Attention Based 3D Channel Estimation

Xiangzhao Qin, Sha Hu

Abstract

For multi-input and multi-output (MIMO) channels, the optimal channel estimation (CE) based on linear minimum mean square error (LMMSE) requires three-dimensional (3D) filtering. However, the complexity is often prohibitive due to large matrix dimensions. Suboptimal estimators approximate 3DCE by decomposing it into time, frequency, and spatial domains, while yields noticeable performance degradation under correlated MIMO channels. On the other hand, recent advances in deep learning (DL) can explore channel correlations in all domains via attention mechanisms. Building on this capability, we propose a dual attention mechanism based 3DCE network (3DCENet) that can achieve accurate estimates.

Dual-Attention Based 3D Channel Estimation

Abstract

For multi-input and multi-output (MIMO) channels, the optimal channel estimation (CE) based on linear minimum mean square error (LMMSE) requires three-dimensional (3D) filtering. However, the complexity is often prohibitive due to large matrix dimensions. Suboptimal estimators approximate 3DCE by decomposing it into time, frequency, and spatial domains, while yields noticeable performance degradation under correlated MIMO channels. On the other hand, recent advances in deep learning (DL) can explore channel correlations in all domains via attention mechanisms. Building on this capability, we propose a dual attention mechanism based 3DCE network (3DCENet) that can achieve accurate estimates.

Paper Structure

This paper contains 15 sections, 28 equations, 6 figures, 1 table.

Table of Contents

  1. Introduction
  2. CE for MIMO System
  3. Correlation Models in Different Domains
  4. The Optimal 3DCE
  5. Suboptimal (2D+1D)-CE and (3$\times$1D)-CE
  6. Optimal Noise Power Allocation
  7. 3DCENet Based on Dual Attention
  8. Preprocess of LS Estimates
  9. Spatial Attention (SA)
  10. Time-Frequency Attention (TFA)
  11. Complexity of SA and TFA Modules
  12. SA
  13. TFA.
  14. Numerical Simulation
  15. Conclusions

Figures (6)

  • Figure 1: An AI receiver design comprised of AI based CE, MIMO detection, and log-likelihood ratios (LLRs) computations QW24ZQ24.
  • Figure 2: DMRS configurations for $N_{t}\!=\!4$, $N_f\!=\!12$, $N_s\!=\!14$, and $N_p\!=\!3$.
  • Figure 3: The proposed 3DCE design that comprises a two-stage process of SA followed by TFA modules.
  • Figure 4: CE-MSEs with different noise power allocations under correlated $4\!\times\!4$ MIMO and ETU100Hz channel.
  • Figure 5: CE-MSEs with the proposed 3DCENet and other methods.
  • ...and 1 more figures