Deep Learning-based Automatic Modulation Recognition Algorithm in Non-Cooperative Communication systems

Automatic modulation recognition (AMR) is considered one of most important topics in non-cooperative communication systems. Traditional algorithms, e.g., high order cumulants (HOC) and support vector machine (SVM), are hard to achieve reliable performance. In this paper, we propose an effective AMR algorithm based on deep learning (DL) with capabilities of automatically extracting representative and effective features. The proposed method resorts to in-phase and quadrature (IQ) samples that are in-phase and quadrature components of received baseband signal, respectively. We adopt convolutional neural networks (CNN) and recurrent neural networks (RNN) to classify six types of signal modulations in additive white Gaussian noise (AWGN) channel and Rayleigh channel, respectively. Simulation results show that the features automatically extracted by DL algorithm is more suitable and effective for AMR and its performance is far beyond that of traditional algorithms in two channels.