/ARIL

Codes of paper: Joint Activity Recognition and Indoor Localization with WiFi Fingerprints

Primary LanguageC

Action Recognition and Indoor Localization

Code and Data of the paper, Joint Activity Recognition and Indoor Localization with WiFi Fingerprints.

Evaluated Environment

  1. PyTorch 1.0.0

Usage

  1. Please download data, and decompress it at the root folder of this repository.

Activity Label: 0. hand up; 1. hand down; 2. hand left; 3. hand right; 4. hand circle; 5. hand cross. Location Label: 0, 1, 2, ..., 15

  1. Please download pre-trained weights, and decompress it at the root folder of this repository.

  2. Then run train.py or test.py

You may need original data (not segmented and upsampled) for your research, here

Hardware: Ettus N210 and Ettus Clock

hardware

1D CNN

  1. 1D residual block

  1. 1D ResNet-[1,1,1,1]

For t-SNE visualization

Please download vis, and run main_plot_tsne.m

  1. t-SNE visualization for activity recognition tsne_act

  2. t-SNE visualization for indoor localization tsne_loc

If this helps your research, please cite this paper.

@article{wang2019joint,
  title={Joint Activity Recognition and Indoor Localization With WiFi Fingerprints},
  author={Wang, Fei and Feng, Jianwei and Zhao, Yinliang and Zhang, Xiaobin and Zhang, Shiyuan and Han, Jinsong},
  journal={IEEE Access},
  volume={7},
  pages={80058--80068},
  year={2019},
}