Skip to content

Latest commit

 

History

History
74 lines (64 loc) · 3.38 KB

README.md

File metadata and controls

74 lines (64 loc) · 3.38 KB
Raw

Long-Lasting UAV-aided RIS Communications based on SWIPT

Introduction

  • This repository is the implementation of "Long-Lasting UAV-aided RIS Communications based on SWIPT" in 2022 IEEE Wireless Communications and Networking Conference (WCNC). [Paper] [Slides] [Video]
  • The implementation of DDPG is based on this tutorial.
  • The architecture of actor and critic nets are implemented via using a MLP (3 layers of 64 neurals).

There are some limitations to this work. If you have any questions or suggestions, please feel free to contact me. Your suggestions are greatly appreciated.

Our extension of this work maximizes the energy harvesting (EH) efficiency by jointly optimizing the phase shift of the reflecting elements, the transmit power, the length of the EH phase, and the RIS scheduling. Please find the latest repository by the below link:
UAV—RIS_EnergyHarvesting for IEEE TWC paper

Citing

Please consider citing our paper if this repository is helpful to you. Bibtex:

@INPROCEEDINGS{peng1570767WCNC,
  author={Peng, Haoran and Wang, Li-Chun and Li, Geoffrey Ye and Tsai, Ang-Hsun},
  booktitle={Proc. IEEE Wireless Commun. Netw. Conf. (WCNC)}, 
  title={Long-Lasting {UAV}-aided {RIS} Communications based on {SWIPT}},
  address={Austin, TX},
  year={2022},
  month = {Apr.}
}

  @ARTICLE{10051712,
  author={Peng, Haoran and Wang, Li-Chun},
  journal={IEEE Trans. Wireless Commun.}, 
  title={Energy Harvesting Reconfigurable Intelligent Surface for {UAV} Based on Robust Deep Reinforcement Learning}, 
  note={early access, Feb. 2023, doi:\url{10.1109/TWC.2023.3245820}}
 }

Requirements

  • Python: 3.6.13
  • Pytorch: 1.10.1
  • gym: 0.15.3
  • numpy: 1.19.2
  • matplotlib
  • pandas

Usage

Descriptions of folders

  • The folder "DDPG-SingleUT-Time" is the source code for the time-domain EH scheme using DDPG.
  • The folder "DDPG-SingleUT-Time-and-Space" is the source code for the two-domain (Time and Space) EH scheme using DDPG.
  • The folder "Exhaustive-SingleUT-Time" is the source code for the time-domain EH scheme using Exhaustive Algorithm.
  • The folder "Exhaustive-SingleUT-Time-and-Space" is the source code for the two-domain (Time and Space) EH scheme using Exhaustive Algorithm.

Descriptions of files

  • For the Exhaustive Algorithm, the communication environment is impletemented in 'ARIS_ENV.py'.
  • For the DDPG, the communication environment is impletemented in 'gym_foo/envs/foo_env.py'.
  • You can change the dataset in 'gym_foo/envs/foo_env.py'.

Training phase

  1. In the main.py, the switch of "Train" must be 'True' such as
14 Train = True # True for tranining, and False for testing.
  1. python main.py

Testing phase

  1. In the main.py, the switch of "Train" must be 'False' such as
14 Train = False # True for tranining, and False for testing.
  1. python main.py
  2. The harvest energy of each step and overall steps are saved in 'Test_Rewards_Records.csv' and 'Total_Reward_Test.txt', respectively.

The EH efficiency

The EH efficiency for each step can be calculated by:

 reward of each step / 0.02275827153828275