This repository provides a code example for end-to-end learning from spectrum data.
End-to-end learning refers to processing architectures where the entire pipeline, connecting the input to the desired output, is learned purely from data.
End-to-end learning from spectrum data is a unified approach that can tackle various challenges related to the problems of inefficient spectrum management, utilization and regulation that the next generation of wireless communication systems are facing. Whether the goal is to recognize a technology or a particular modulation type, identify the interference source or an interference-free frequency channel, we argue that the various problems may be treated as a generic problem type that we refer to as RF signal identification for which a unified solution applies.
The code provides modules for setting up and training a CNN network, and collecting performance evaluation results for wireless signal modulation recognition.
Should you find this helpful please cite our work:
Kulin, M., Kazaz, T., Moerman, I., & De Poorter, E. (2018). End-to-end learning from spectrum data: A deep learning approach for wireless signal identification in spectrum monitoring applications. IEEE Access, 6, 18484-18501.
Or
@article{kulin2018end,
title={End-to-end learning from spectrum data: A deep learning approach for wireless signal identification in spectrum monitoring applications},
author={Kulin, Merima and Kazaz, Tarik and Moerman, Ingrid and De Poorter, Eli},
journal={IEEE Access},
volume={6},
pages={18484--18501},
year={2018},
publisher={IEEE}
}