Deep Neural Network Architectures for Modulation Classification

This thesis investigates the value of employing deep learning for the task of wireless signal modulation recognition. Recently in deep learning research on AMC, a framework has been introduced by generating a dataset using GNU radio that mimics the imperfections in a real wireless channel, and uses 10 different modulation types. Further, a CNN architecture was developed and shown to deliver performance that exceeds that of expert-based approaches. Here, we follow the framework of O’shea [1] and find deep neural network architectures that deliver higher accuracy than the state of the art. We tested the architecture of O’shea [1] and found it to achieve an accuracy of approximately 75% of correctly recognizing the modulation type. We first tune the CNN architecture and find a design with four convolutional layers and two dense layers that gives an accuracy of approximately 83.8% at high SNR. We then develop architectures based on the recently introduced ideas of Residual Networks (ResNet) and Densely Connected Network (DenseNet) to achieve high SNR accuracies of approximately 83% and 86.6%, respectively. We also introduce a CLDNN to achieve an accuracy of approximately 88.5% at high SNR. To improve the classification accuracy of QAM, we calculate the high order cumulants of QAM16 and QAM64 as the expert feature and improve the total accuracy to approximately 90%. Finally, by preprocessing the input and send them into a LSTM model, we improve all classification success rates to 100% except the WBFM which is 46%. The average modulation classification accuracy got a improvement of roughly 22% in this thesis