WRF-GS: Wireless Radiation Field Reconstruction with 3D Gaussian Splatting

Chaozheng Wen, Jingwen Tong, Yingdong Hu, Zehong Lin and Jun Zhang

Thanks for your interest in our work. This repository contains the official implementation associated with the paper " WRF-GS: Wireless Radiation Field Reconstruction with 3D Gaussian Splatting", which has been accepted for INFOCOM 2025 ArXiv.

News

• [4/11/2025] WRF-GS+ has been accepted by TWC 2025.
• [3/24/2025] An enhanced version WRF-GS+ has been released.

Abstract: Wireless channel modeling plays a pivotal role in designing, analyzing, and optimizing wireless communication systems. Nevertheless, developing an effective channel modeling approach has been a longstanding challenge. This issue has been escalated due to the denser network deployment, larger antenna arrays, and wider bandwidth in 5G and beyond networks. To address this challenge, we put forth WRF-GS, a novel framework for channel modeling based on wireless radiation field (WRF) reconstruction using 3D Gaussian splatting. WRF-GS employs 3D Gaussian primitives and neural networks to capture the interactions between the environment and radio signals, enabling efficient WRF reconstruction and visualization of the propagation characteristics. The reconstructed WRF can then be used to synthesize the spatial spectrum for comprehensive wireless channel characterization. Notably, with a small number of measurements, WRF-GS can synthesize new spatial spectrums within milliseconds for a given scene, thereby enabling latency-sensitive applications. Experimental results demonstrate that WRF-GS outperforms existing methods for spatial spectrum synthesis, such as ray tracing and other deep-learning approaches.

Installation

Create the basic environment

conda env create --file environment.yml
conda activate wrfgs

Install some extensions

cd submodules
pip install ./simple-knn
pip install ./diff-gaussian-rasterization # or cd ./diff-gaussian-rasterization && python setup.py develop
pip install ./submodules/fused-ssim

Training & Evaluation

Due to file size limitations, a small dataset is included to help quickly verify the code, which can be executed using the following command:

python train.py

More datasets can be found here.

Note: As the amount of training data changes, you may need to adjust the total iterations (./arguments/__init__.py) and learning rate (./scene/gaussian_model.py).

To-Do List

• Implementation on gsplat to speed up the code.
• Release more case study code.
• Optimized code is applicable to more scenarios.

BibTex

If you find this work useful in your research, please cite:

@INPROCEEDINGS{11044513,
  author={Wen, Chaozheng and Tong, Jingwen and Hu, Yingdong and Lin, Zehong and Zhang, Jun},
  booktitle={IEEE INFOCOM 2025 - IEEE Conference on Computer Communications}, 
  title={WRF-GS: Wireless Radiation Field Reconstruction with 3D Gaussian Splatting}, 
  year={2025},
  volume={},
  number={},
  pages={1-10},
  keywords={Wireless communication;Training;Solid modeling;Visualization;Adaptation models;Three-dimensional displays;Neural networks;Ray tracing;Rendering (computer graphics);Antenna arrays;Wireless channel modeling;wireless radiation field reconstruction;3D Gaussian splatting;channel prediction},

  doi={10.1109/INFOCOM55648.2025.11044513}}

Acknowledgment

Thanks to the NeRF2 author for his help. And some code snippets are borrowed from 3DGS.