Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Pay Less But Get More: A Dual-Attention-based Channel Estimation Network for Massive MIMO Systems with Low-Density Pilots

Binggui Zhou, Xi Yang, Shaodan Ma, Feifei Gao, Guanghua Yang

TL;DR

A dual-attention-based channel estimation network (DACEN) is proposed to realize accurate channel estimation via low-density pilots, by jointly learning the spatial-temporal domain features of massive MIMO channels with the temporal attention module and the spatial attention module.

Abstract

To reap the promising benefits of massive multiple-input multiple-output (MIMO) systems, accurate channel state information (CSI) is required through channel estimation. However, due to the complicated wireless propagation environment and large-scale antenna arrays, precise channel estimation for massive MIMO systems is significantly challenging and costs an enormous training overhead. Considerable time-frequency resources are consumed to acquire sufficient accuracy of CSI, which thus severely degrades systems' spectral and energy efficiencies. In this paper, we propose a dual-attention-based channel estimation network (DACEN) to realize accurate channel estimation via low-density pilots, by jointly learning the spatial-temporal domain features of massive MIMO channels with the temporal attention module and the spatial attention module. To further improve the estimation accuracy, we propose a parameter-instance transfer learning approach to transfer the channel knowledge learned from the high-density pilots pre-acquired during the training dataset collection period. Experimental results reveal that the proposed DACEN-based method achieves better channel estimation performance than the existing methods under various pilot-density settings and signal-to-noise ratios. Additionally, with the proposed parameter-instance transfer learning approach, the DACEN-based method achieves additional performance gain, thereby further demonstrating the effectiveness and superiority of the proposed method.

Pay Less But Get More: A Dual-Attention-based Channel Estimation Network for Massive MIMO Systems with Low-Density Pilots

TL;DR

A dual-attention-based channel estimation network (DACEN) is proposed to realize accurate channel estimation via low-density pilots, by jointly learning the spatial-temporal domain features of massive MIMO channels with the temporal attention module and the spatial attention module.

Abstract

To reap the promising benefits of massive multiple-input multiple-output (MIMO) systems, accurate channel state information (CSI) is required through channel estimation. However, due to the complicated wireless propagation environment and large-scale antenna arrays, precise channel estimation for massive MIMO systems is significantly challenging and costs an enormous training overhead. Considerable time-frequency resources are consumed to acquire sufficient accuracy of CSI, which thus severely degrades systems' spectral and energy efficiencies. In this paper, we propose a dual-attention-based channel estimation network (DACEN) to realize accurate channel estimation via low-density pilots, by jointly learning the spatial-temporal domain features of massive MIMO channels with the temporal attention module and the spatial attention module. To further improve the estimation accuracy, we propose a parameter-instance transfer learning approach to transfer the channel knowledge learned from the high-density pilots pre-acquired during the training dataset collection period. Experimental results reveal that the proposed DACEN-based method achieves better channel estimation performance than the existing methods under various pilot-density settings and signal-to-noise ratios. Additionally, with the proposed parameter-instance transfer learning approach, the DACEN-based method achieves additional performance gain, thereby further demonstrating the effectiveness and superiority of the proposed method.
Paper Structure (19 sections, 30 equations, 9 figures, 6 tables, 1 algorithm)

This paper contains 19 sections, 30 equations, 9 figures, 6 tables, 1 algorithm.

Table of Contents

  1. Introduction
  2. System Model and Problem Formulation
  3. System Model
  4. Problem Formulation
  5. Pilot Density
  6. Dual-Attention Channel Estimation Network
  7. Overall Network Architecture
  8. Temporal Attention Module
  9. Spatial Attention Module
  10. Learning Strategy
  11. Parameter-Instance Transfer Learning for Low-overhead Channel Estimation
  12. Transfer Learning
  13. Proposed Parameter-Instance Transfer Learning Approach
  14. Experiments
  15. Experimental Settings
  16. ...and 4 more sections

Figures (9)

  • Figure 1: An illustration of a spatial-temporal channel representation.
  • Figure 2: Proposed dual-attention channel estimation network (DACEN). (a). The overall network architecture of the DACEN. (b). The spatial attention mechanism. (c). The temporal attention mechanism. RC: residual connection. LN: layer normalization. PE: positional encoding. C: concatenation operation. R: rearrangement operation.
  • Figure 3: Training and testing processes of the DACEN. (a). Training the DACEN from scratch. (b) Training the DACEN with the parameter-instance transfer learning approach. (c). Testing the DACEN with low-density pilots.
  • Figure 4: The CSI-RS configuration within one RB.
  • Figure 5: NMSE performance under different SNRs with $\rho_H=26/52$.
  • ...and 4 more figures