Cross-Person HAR

The content of this article was published in IEEE VTC2022-Fall

Dataset

1. Download dataset from here:

   Link: https://pan.baidu.com/s/1Tgwd8CIDaWfFCCbfrRX8DQ?pwd=axhz

2. A dataset for seven different daily human activities including wave, clap, walk, liedown, sitdown, fall and pickup in an indoor environment.

3. We use ESP32 CSI Tool to collect CSV files in our dataset.

Code

1. SE-ABLSTM-trainmodels.py: Train ten models at once with cosine annealing learning rate
2. SE-ABLSTM-test.py: Use previously generated models to make ensemble predictions on datasets of different people

Reference

1. https://github.com/ermongroup/Wifi_Activity_Recognition

2. https://github.com/parisafm/CSI-HAR-Dataset

3. https://github.com/ludlows/CSI-Activity-Recognition

4. https://github.com/StevenMHernandez/ESP32-CSI-Tool