Unsupervised Representation Learning of Structured Radio Communication Signals
This work addresses the problem of signal representation for radio communication analysis, but it appears incremental as it applies existing autoencoder methods to a specific domain without claiming major breakthroughs.
The paper tackled unsupervised representation learning of radio communication signals from raw time series data, demonstrating that convolutional autoencoders can learn modulation basis functions and showing their relationship to analytic bases used in digital communications.
We explore unsupervised representation learning of radio communication signals in raw sampled time series representation. We demonstrate that we can learn modulation basis functions using convolutional autoencoders and visually recognize their relationship to the analytic bases used in digital communications. We also propose and evaluate quantitative met- rics for quality of encoding using domain relevant performance metrics.