Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Adaptive TTD Configurations for Near-Field Communications: An Unsupervised Transformer Approach

Hsienchih Ting, Zhaolin Wang, Yuanwei Liu

TL;DR

This work tackles near-field wideband beamforming in XL-MIMO by proposing an adaptive-serial TTD architecture that dynamically connects TTDs to PSs via a switch network. A novel unsupervised end-to-end framework, comprising the NFC-LM (UNet-based channel learning with cross-attention) and the S-MT (switch Multi-User Transformer with Hungarian assignment), jointly optimizes analog beamformers and switch connections, enhanced further by the MCA module. The approach achieves near-optimal spectral efficiency across arbitrary user locations and array shapes (ULA and UCA), reduces hardware complexity, and enables real-time adaptation. The results demonstrate strong mitigation of the spatial-wideband effect under short-range TTDs and across multi-user scenarios, suggesting significant practical impact for 6G/6G+ high-frequency deployments with flexible hardware constraints.

Abstract

True-time delayers (TTDs) are popular analog devices for facilitating near-field wideband beamforming subject to the spatial-wideband effect. In this paper, an adaptive TTD configuration is proposed for short-range TTDs. Compared to the existing TTD configurations, the proposed one can effectively combat the spatial-widebandd effect for arbitrary user locations and array shapes with the aid of a switch network. A novel end-to-end deep neural network is proposed to optimize the hybrid beamforming with adaptive TTDs for maximizing spectral efficiency. 1) First, based on the U-Net architecture, a near-field channel learning module (NFC-LM) is proposed for adaptive beamformer design through extracting the latent channel response features of various users across different frequencies. In the NFC-LM, an improved cross attention (CA) is introduced to further optimize beamformer design by enhancing the latent feature connection between near-field channel and different beamformers. 2) Second, a switch multi-user transformer (S-MT) is proposed to adaptively control the connection between TTDs and phase shifters (PSs). In the S-MT, an improved multi-head attention, namely multi-user attention (MSA), is introduced to optimize the switch network through exploring the latent channel relations among various users. 3) Third, a multi feature cross attention (MCA) is introduced to simultaneously optimize the NFC-LM and S-MT by enhancing the latent feature correlation between beamformers and switch network. Numerical simulation results show that 1) the proposed adaptive TTD configuration effectively eliminates the spatial-wideband effect under uniform linear array (ULA) and uniform circular array (UCA) architectures, and 2) the proposed deep neural network can provide near optimal spectral efficiency, and solve the multi-user bemformer design and dynamical connection problem in real-time.

Adaptive TTD Configurations for Near-Field Communications: An Unsupervised Transformer Approach

TL;DR

This work tackles near-field wideband beamforming in XL-MIMO by proposing an adaptive-serial TTD architecture that dynamically connects TTDs to PSs via a switch network. A novel unsupervised end-to-end framework, comprising the NFC-LM (UNet-based channel learning with cross-attention) and the S-MT (switch Multi-User Transformer with Hungarian assignment), jointly optimizes analog beamformers and switch connections, enhanced further by the MCA module. The approach achieves near-optimal spectral efficiency across arbitrary user locations and array shapes (ULA and UCA), reduces hardware complexity, and enables real-time adaptation. The results demonstrate strong mitigation of the spatial-wideband effect under short-range TTDs and across multi-user scenarios, suggesting significant practical impact for 6G/6G+ high-frequency deployments with flexible hardware constraints.

Abstract

True-time delayers (TTDs) are popular analog devices for facilitating near-field wideband beamforming subject to the spatial-wideband effect. In this paper, an adaptive TTD configuration is proposed for short-range TTDs. Compared to the existing TTD configurations, the proposed one can effectively combat the spatial-widebandd effect for arbitrary user locations and array shapes with the aid of a switch network. A novel end-to-end deep neural network is proposed to optimize the hybrid beamforming with adaptive TTDs for maximizing spectral efficiency. 1) First, based on the U-Net architecture, a near-field channel learning module (NFC-LM) is proposed for adaptive beamformer design through extracting the latent channel response features of various users across different frequencies. In the NFC-LM, an improved cross attention (CA) is introduced to further optimize beamformer design by enhancing the latent feature connection between near-field channel and different beamformers. 2) Second, a switch multi-user transformer (S-MT) is proposed to adaptively control the connection between TTDs and phase shifters (PSs). In the S-MT, an improved multi-head attention, namely multi-user attention (MSA), is introduced to optimize the switch network through exploring the latent channel relations among various users. 3) Third, a multi feature cross attention (MCA) is introduced to simultaneously optimize the NFC-LM and S-MT by enhancing the latent feature correlation between beamformers and switch network. Numerical simulation results show that 1) the proposed adaptive TTD configuration effectively eliminates the spatial-wideband effect under uniform linear array (ULA) and uniform circular array (UCA) architectures, and 2) the proposed deep neural network can provide near optimal spectral efficiency, and solve the multi-user bemformer design and dynamical connection problem in real-time.
Paper Structure (22 sections, 38 equations, 11 figures, 2 tables, 1 algorithm)

This paper contains 22 sections, 38 equations, 11 figures, 2 tables, 1 algorithm.

Table of Contents

  1. Introduction
  2. Prior Works
  3. Conventional Algorithm based Hybrid Beamforming
  4. Deep Learning based Hybrid Beamforming
  5. Motivation and Contributions
  6. Organization and Notations
  7. System Model and Problem Formulation
  8. Signal Model
  9. Proposed Adaptive-Serial TTD Configuration
  10. Problem Formulation
  11. DL-Based Adaptive TTD Hybrid Beamforming
  12. Near-Field Channel Latent Feature Learning
  13. Switch Multi-User Transformer
  14. Network Architecture for Adaptive TTD Beamforming
  15. Loss Function and Network Training Process
  16. ...and 7 more sections

Figures (11)

  • Figure 1: Proposed adaptive-serial configuration for TTD-based hybrid beamforming.
  • Figure 2: The network structure of channel feature learning module.
  • Figure 3: The network structure of multi-user transformer for adaptive connection between TTDs and PSs.
  • Figure 4: The network structure of the proposed adaptive TTD configuration beamforming method.
  • Figure 5: Spectral efficiency results of different network structures based on the NFC-LM across varying levels of transmit power.
  • ...and 6 more figures