Multicarrier waveforms classification with deep neural networks

Automatic modulation classification (AMC) plays an important role in various applications such as cognitive radio and dynamic spectrum access. Many research works have been exploring deep learning (DL) based AMC, but they primarily focus on single-carrier signals. With the advent of various multicarrier waveforms, the authors propose to revisit DL-based AMC to consider the diversity and complexity of these novel transmission waveforms in this letter. Specifically, the authors develop a novel representation of multicarrier signals and use suitable networks for classification. In addition, to cope with non-target signals, support vector data description (SVDD) is applied with the activations of the networks’ hidden layer. Experimental results demonstrate the effectiveness of the proposed scheme.