Nicola Novello and Andrea M. Tonello
Official repository of the paper "Recurrent DQN for Radio Fingerprinting with Constrained Measurements Collection" published at ICT Express (2024).
The folder where the code is located must contain a folder NetsImages
comprising the following folders:
Datasets
containing the RSSI dataframe obtained after interpolation of the dataset in https://github.com/beaugunderson/wifi-heatmap and the initial training datasets randomly extracted from the same datasetDRQN
containing the saved DRQNFloorImages
containing the map layoutOptimalPaths
containing the images of the saved optimal paths traveled by the agentRegModels
containing the saved PPNNsscalers
containing the scaler used to standardize the datasetaccuraciesComparison
containing the saved test accuracies
The file main.py
runs the experiments, while main_functions.py
, classes.py
, and utils.py
comprise the classes and functions needed to run the main script.
If you use your code for your research, please cite our paper:
@article{novello2024recurrent,
title={Recurrent DQN for radio fingerprinting with constrained measurements collection},
author={Novello, Nicola and Tonello, Andrea M},
journal={ICT Express},
year={2024},
publisher={Elsevier}
}
The implementation is based on / inspired by:
- https://github.com/beaugunderson/wifi-heatmap
- https://mlpeschl.com/post/tiny_adrqn/
- https://github.com/patrickloeber/snake-ai-pytorch