Lzy-ee / testSpectrogram

testSpectrogram is an open-source platform for wireless channel simulation, human/hand pose extraction, gesture spectrogram generation, and real-time gesture recognition based on millimeter-wave passive sensing and communication systems.

testSpectrogram

See THREE video demos at https://lasso-sustech.github.io/CASTER or http://lasso.eee.sustech.edu.cn/caster/ for a quick understanding of our efforts!

SDP3 (IOTJ)	PBAH (SPAWC)	CASTER (OJ-COMS)

Introduction

testSpectrogram is an open-source platform for wireless channel simulation, human/hand pose extraction, gesture spectrogram generation, and real-time gesture recognition based on millimeter-wave passive sensing and communication systems.

Since this work is experiment-oriented, code might not be 100% consistent with our real implementation. But the gerneral core idea is the same. And we will keep making improvements and updating the code for clearer understanding.

Code Overview

Clone the Repository

git clone https://github.com/rzy0901/testSpectrogram.git --recursive

Alternatively, you can visit the repositories listed in .gitmodules and download each one individually as a zip file.

Code for "SDP3" and "PBAH" papers

• Micro_Doppler_Radar_Simulator
  • Data driven hybrid channel model simulation using a Boulic Human walking model.
• testZED and zed_pose
  • Simple Mocap-based channel simulation example.
  • Camera coordinate 3D human keypoints extraction based on the depth camera ZED 2i, using zed-sdk.

Code for "CASTER" paper

• mediapipe_spectrogram
  • Primitive-based wireless channel simulation for hand gesture recognition.
  • Camera coordinate 3D hand keypoints extraction based on a monocular camera, using mediapipe and opencv.
• CASTER_classification and RxRealTime_GUI_rzy
  • "Simulation-to-reality" hand gesture recognition based on ResNet18.
  • Transfer learning based on the simulated dataset and real-world dataset.
  • Real-time gesture recognition based on millimeter-wave passive sensing and communication systems, using a model trained by a simulated dataset.

Cite this repository

@inproceedings{li2021wireless,
  title={Wireless sensing with deep spectrogram network and primitive based autoregressive hybrid channel model},
  author={Li, Guoliang and Wang, Shuai and Li, Jie and Wang, Rui and Peng, Xiaohui and Han, Tony Xiao},
  booktitle={2021 IEEE 22nd International Workshop on Signal Processing Advances in Wireless Communications (SPAWC)},
  pages={481--485},
  year={2021},
  organization={IEEE}
}

@article{li2023integrated,
  title={Integrated Sensing and Communication from Learning Perspective: An SDP3 Approach},
  author={Li, Guoliang and Wang, Shuai and Li, Jie and Wang, Rui and Liu, Fan and Peng, Xiaohui and Han, Tony Xiao and Xu, Chengzhong},
  journal={IEEE Internet of Things Journal},
  year={2023},
  publisher={IEEE}
}

@ARTICLE{ren2024caster,
  author={Ren, Zhenyu and Li, Guoliang and Ji, Chenqing and Yu, Chao and Wang, Shuai and Wang, Rui},
  journal={IEEE Open Journal of the Communications Society}, 
  title={CASTER: A Computer-Vision-Assisted Wireless Channel Simulator for Gesture Recognition}, 
  year={2024},
  volume={5},
  number={},
  pages={3185-3195},
  doi={10.1109/OJCOMS.2024.3398016},
  ISSN={2644-125X},
  month={},}

Authors

Zhenyu Ren

Guoliang Li

Shuai Wang

Chenqing Ji

Chao Yu

Acknowledgement

This series of work is under supervision of Prof. Rui Wang and Prof. Shuai Wang.