Wideband Beamforming for RIS Assisted Near-Field Communications

We highly respect reproducible research, so we try to provide the simulation codes for our submitted papers. Please refer to the following paper for more details.

@ARTICLE{10659326, author={Wang, Ji and Xiao, Jian and Zou, Yixuan and Xie, Wenwu and Liu, Yuanwei},
journal={IEEE Transactions on Wireless Communications} ,
title={Wideband Beamforming for RIS Assisted Near-Field Communications},
year={2024},
doi={10.1109/TWC.2024.3447570}}

Script descriptions.

• main_RIS_WB_SNR.py: the main function of the proposed E2E model.
  
• RIS_SUB_DIR_MIMO_NFWB_SNR.py: the SA-RIS architecture.
  
• RIS_TDD_DIR_MIMO_NFWB_SNR.py: the TTD-RIS architecture.
  
• RIS_SUB_DIR_MIMO_NFWB_SNR_R1.py: the SA-RIS architecture with the quantified phase shift.
  
• RIS_TDD_DIR_MIMO_NFWB_SNR_R1.py: the TTD-RIS architecture with the quantified phase shift.
  
• PolarizedSelfAttention.py: polarized attention module.
  
• GFNet.py: learnable DFT module.
  
• Transformer_model.py: transformer module.
  
• NFBF_RIS_R3.py: channel generation functions.
  

Usage

1. You can run the “main_RIS_WB_SNR.py” script to obtain the desired results by switching different beamforming models.
   
2. You can run the "NFBF_RIS_R3.py" script to generate a common test dataset at first so that provides a fairness performance comparison among various schemes. In this case, the system parameters between "NFBF_RIS_R3.py" and “main_RIS_WB_SNR.py” scripts must be consistent.
   
3. When you call the "RIS_SUB_DIR_MIMO_NFWB_SNR_R1.py" script for evaluating the beamforming performance under the case of discrete phase shift, you should pretrain an infinite beamforming model with the "RIS_SUB_DIR_MIMO_NFWB_SNR_R1.py" script at first.
   
4. In the training stage, the different hyper-parameters will result in slight difference for final beamforming performance, e.g., the batchsize, the number of training epochs, and the training learning rate.
   
5. Now, this code is a preliminary version composed of a few redundant statements, we will try my best to release the clean codes and add the necessary annotations in the future.
   

Acknowledgements

We are very grateful for the following open-source repositories, which help us construct the beamforming model.

1. https://github.com/DeLightCMU/PSA
   
2. https://github.com/raoyongming/GFNet
   
3. https://github.com/wuminghui123/DL_RSMA