Robust Wifi-enabled Device-free Gesture recognition via Unsupervised Adversarial Domain Adaptation

Introduction

In this paper, we propose a WiFi-enabled device-free adaptive gesture recognition scheme, WiADG, that is able to identify human gestures accurately and consistently under environmental dynamics via adversarial domain adaptation. (link)

Requirements

• Python3
• Pytorch (1.1.0 Recommended)

Guidance

• Train the source encoder and classifier

python train_src.py

• Test the source on the target dataset

python test_src.py

• Train the target encoder

python train_adapt.py

• Test the target encoder with source classifier on the target dataset

python test_tgt.py

Performance

Method	Conf room --> Office	Office --> Conf room
Source-only	49.7%	32.7%
Target-only	96.7%	93.0%
WiADG (Ours)	83.3%	66.6%

Reference

@inproceedings{zou2018robust,
  title={Robust wifi-enabled device-free gesture recognition via unsupervised adversarial domain adaptation},
  author={Zou, Han and Yang, Jianfei and Zhou, Yuxun and Xie, Lihua and Spanos, Costas J},
  booktitle={2018 27th International Conference on Computer Communication and Networks (ICCCN)},
  pages={1--8},
  year={2018},
  organization={IEEE},
  doi={10.1109/ICCCN.2018.8487345}
}