Deep learning solutions for large scale wireless spectrum monitoring

Sreeraj Rajendrnan

KU Leuven

Wireless spectrum usage knowledge is critical for various applications ranging from spectrum enforcement to dynamic spectrum sharing. This talk will cover new steps towards the development of a large scale wireless spectrum monitoring and analysis framework leveraging the recent advancements in machine learning, especially some of the deep learning models. We achieve the spectrum monitoring goal by answering the following four major research questions. 1) How to monitor the complex and dense electromagnetic space on a large scale in a cost-effective manner? 2) How to detect and classify anomalous behaviors in the wireless spectrum efficiently without supervision? 3) How to incorporate user feedback to the entire monitoring framework for improving spectrum awareness? 4) How good is state-of-the-art research in utilizing the achieved spectrum awareness for improving wireless communication performance? The pros and cons of the newly proposed crowdsourced spectrum monitoring, anomaly detection and signal classification models will be covered during the talk.