Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Multimodal-NF: A Wireless Dataset for Near-Field Low-Altitude Sensing and Communications

Mengyuan Li, Qianfan Lu, Jiachen Tian, Hongjun Hu, Yu Han, Xiao Li, Chao-Kai Wen, Shi Jin

Abstract

Environment-aware 6G wireless networks demand the deep integration of multimodal and wireless data. However, most existing datasets are confined to 2D terrestrial far-field scenarios, lacking the 3D spatial context and near-field characteristics crucial for low-altitude extremely large-scale multiple-input multiple-output (XL-MIMO) systems. To bridge this gap, this letter introduces Multimodal-NF, a large-scale dataset and specialized generation framework. Operating in the upper midband, it synchronizes high-fidelity near-field channel state information (CSI) and precise wireless labels (e.g., Top-5 beam indices, LoS/NLoS) with comprehensive sensory modalities (RGB images, LiDAR point clouds, and GPS). Crucially, these multimodal priors provide spatial semantics that help reduce the near-field search space and thereby lower the overhead of wireless sensing and communication tasks. Finally, we validate the dataset through representative case studies, demonstrating its utility and effectiveness. The open-source generator and dataset are available at https://lmyxxn.github.io/6GXLMIMODatasets/.

Multimodal-NF: A Wireless Dataset for Near-Field Low-Altitude Sensing and Communications

Abstract

Environment-aware 6G wireless networks demand the deep integration of multimodal and wireless data. However, most existing datasets are confined to 2D terrestrial far-field scenarios, lacking the 3D spatial context and near-field characteristics crucial for low-altitude extremely large-scale multiple-input multiple-output (XL-MIMO) systems. To bridge this gap, this letter introduces Multimodal-NF, a large-scale dataset and specialized generation framework. Operating in the upper midband, it synchronizes high-fidelity near-field channel state information (CSI) and precise wireless labels (e.g., Top-5 beam indices, LoS/NLoS) with comprehensive sensory modalities (RGB images, LiDAR point clouds, and GPS). Crucially, these multimodal priors provide spatial semantics that help reduce the near-field search space and thereby lower the overhead of wireless sensing and communication tasks. Finally, we validate the dataset through representative case studies, demonstrating its utility and effectiveness. The open-source generator and dataset are available at https://lmyxxn.github.io/6GXLMIMODatasets/.

Paper Structure

This paper contains 13 sections, 1 theorem, 10 equations, 6 figures, 4 tables.

Table of Contents

  1. Introduction
  2. System Model
  3. Near-Field Channel Model
  4. Motivation for Multimodal Support
  5. Multimodal-NF: Dataset Generator
  6. Scene Construction and Trajectory Simulation
  7. Wireless and Multimodal Data Generation
  8. Wireless Data
  9. Multimodal Sensing Data
  10. Dataset Analysis and Case Studies
  11. Case Study I: User localization
  12. Case Study II: Multimodal Beam Prediction
  13. Conclusion

Key Result

Proposition 1

If the sensing-related target $\mathbf{s}_t$ and the communication-related target $\mathbf{c}_t$ both depend on the geometric state $\mathbf{x}_t$ up to small residual uncertainties, quantified by For multimodal observation $V_t$, the residual uncertainties satisfy In the special case where both $\mathbf{s}_t$ and $\mathbf{c}_t$ are deterministic functions of $\mathbf{x}_t$, the above bounds red

Figures (6)

  • Figure 1: Illustration of the low-altitude XL-MIMO system.
  • Figure 2: Example visualizations of (a) the LAE scene with the pre-defined trajectory modes and (b) the near-field rays.
  • Figure 3: Visualizations of (a) the Cartesian-domain channel and 3D location of UE (estimation and GT) and (b) the Angular-domain channel and angles of UE (estimation and GT).
  • Figure 4: Spatial-temporal variation of beam indices in (a) LoS and (b) NLoS scenarios, including the global index variations and decomposed distance $d$, elevation $\varphi$, and azimuth $\theta$ indices over time. The color indicates the beam index value.
  • Figure 5: System achievable rate comparison of the LLM-based beam prediction method (trained on the proposed dataset) with traditional beam training baselines.
  • ...and 1 more figures

Theorems & Definitions (2)

  • Proposition 1: Entropy Reduction for Sensing and Communication via Multimodal Side Information
  • Remark 1