Automatic Modulation Classification with Deep Neural Networks
This work addresses the need for efficient modulation classification in software-defined radios, but it is incremental as it builds on existing deep learning methods without introducing a fundamentally new approach.
The paper tackled the problem of automatic modulation classification by conducting a comprehensive ablation study of convolutional neural network architectures, achieving a new state-of-the-art performance with a specific combination of design elements like dilated convolutions, statistics pooling, and squeeze-and-excitation units.
Automatic modulation classification is a desired feature in many modern software-defined radios. In recent years, a number of convolutional deep learning architectures have been proposed for automatically classifying the modulation used on observed signal bursts. However, a comprehensive analysis of these differing architectures and importance of each design element has not been carried out. Thus it is unclear what tradeoffs the differing designs of these convolutional neural networks might have. In this research, we investigate numerous architectures for automatic modulation classification and perform a comprehensive ablation study to investigate the impacts of varying hyperparameters and design elements on automatic modulation classification performance. We show that a new state of the art in performance can be achieved using a subset of the studied design elements. In particular, we show that a combination of dilated convolutions, statistics pooling, and squeeze-and-excitation units results in the strongest performing classifier. We further investigate this best performer according to various other criteria, including short signal bursts, common misclassifications, and performance across differing modulation categories and modes.