• Stars
    star
    426
  • Rank 101,884 (Top 3 %)
  • Language
  • License
    MIT License
  • Created over 2 years ago
  • Updated 4 months ago

Reviews

There are no reviews yet. Be the first to send feedback to the community and the maintainers!

Repository Details

A list of awesome papers and cool resources on WiFi CSI sensing.

GitHub Maintenance Ask Me Anything ! Awesome

Awesome WiFi Sensing

A list of awesome papers and cool resources on WiFi CSI sensing. Link to the Github if available is also present.

You are very welcome to suggest resources via pull requests.

Table of Contents

Benchmark

Papers

Papers are ordered by theme and inside each theme by publication date (submission date for arXiv papers).

Methods

WiFi CSI sensing methods have enabled many applications, which can be divided into three categories:

  • Learning-based methods learn the mapping functions from CSI data to the corresponding labels by machine learning and deep learning.
  • Modeling-based methods are based on physical theories like the Fresnel Zone model, or statistical models like the Rician fading model.
  • Hybrid methods derive the strengths from learning-based and modeling-based methods.

Surveys

Applications

Occupancy Detection

Human Activity Recognition

Human Identification

Crowd Counting

Gesture Recognition

Fall Detection

Vital Sign Detection & Healthcare

In-Car Activity Recognition

Pose Estimation

Indoor Localization

Challenges for Real-World Large-Scale WiFi Sensing

IoT System Design

Efficiency and Security

Cross-Environment WiFi Sensing

Multi-modal Sensing (WiFi+CV/Radar)

Platforms

CSI Tool

Datasets

  • [NTU-Fi] The NTU-Fi dataset is the only CSI dataset with 114 subcarriers per pair of antennas, collected using Atheros CSI tool. It consists of 6 human activities and 14 human gait patterns.
  • [Widar 3.0] The Widar 3.0 project is a large dataset designed for use in WiFi-based hand gesture recognition, collected using Intel 5300 NIC (30 subcarriers). The dataset consists of 258K instances of hand gestures with a duration of totally 8,620 minutes and from 75 domains.
  • [WiAR] The WiAR dataset contains sixteen activities including coarse-grained activity and gestures performed by ten volunteers with 30 times every volunteer.
  • [UT-HAR] The dataset is collected in ''A Survey on Behaviour Recognition Using WiFi Channel State Information''. It consists of continuous CSI data for 6 activities without golden segmentation timestamp for each sample.
  • [SignFi] Channel State Information (CSI) traces for sign language recognition using WiFi.

Libraries & Codes

Libraries

  • [SenseFi] Deep learning libraries for WiFi CSI sensing (PyTorch) (Model Zoo)

Github Repositories

From Papers

From Developers

Book Chapter

Reference

If you think that this github project is helpful for your research, please cite the following paper that includes the above works before 2022.

@article{yang2023benchmark,
  title={SenseFi: A Library and Benchmark on Deep-Learning-Empowered WiFi Human Sensing},
  author={Yang, Jianfei and Chen, Xinyan and Wang, Dazhuo and Zou, Han and Lu, Chris Xiaoxuan and Sun, Sumei and Xie, Lihua},
  journal={Patterns},
  volume={4},
  number={3},
  publisher={Elsevier},
  year={2023}
}