Deep neural network-based blind modulation classification for fading channels
Jung‐Hwan Lee, ByeoungDo Kim, Jaekyum Kim, Dongweon Yoon, Jun Won Choi · 2017
In this paper, we propose high performance blind modulation classification (BMC) technique based on deep neural network (DNN) for fading channels. First, we provide the large and diverse set of the features that exhibit statistical relevance to modulation class in fading channels. Then, we use those features to train the DNN to classify the modulation class. Owing to the capability of DNN to learn the complex structure in high dimensional feature space, the proposed scheme achieves the excellent classification accuracy using a number of features in challenging fading environments. Numerical evaluation demonstrates the superiority of the proposed technique over the existing BMC methods.