Deep Learning Based Automatic Modulation Classification Using IQ Signals

In this study, we explore the application of deep learning techniques for Automatic Modulation Classification (AMC) utilizing In-phase and Quadrature (IQ) signals. Specifically, we leverage the RadioML 2018.01A dataset, which contains various modulation types, to train and evaluate our models. Our approach employs Convolutional Neural Networks (CNNs) to effectively capture both spatial and temporal features inherent in the IQ data. We also experiment with different architectures and hyperparameters to optimize classification performance. The results demonstrate that deep learning methods can achieve high accuracy in identifying modulation schemes, outperforming traditional machine learning techniques. This work highlights the potential of deep learning in enhancing the capabilities of wireless communication systems through improved signal processing and classification.