Deep Learning-assisted Automatic Modulation Classification using Spectrograms

With the increasing demand for reliable and efficient V2X (Vehicle-to-Everything) communications in cognitive radio environments, spectrum sharing becomes imperative. In this context, accurate modulation classification serves as a fundamental component for efficient spectrum sensing and allocation. This paper proposes a novel approach utilizing Convolutional Neural Networks (CNNs) trained on spectrograms of BPSK and QPSK modulation schemes for automatic modulation classification in V2X scenarios. Experimental results demonstrated the effectiveness of the proposed CNN-based framework in accurately classifying modulation schemes in V2X communications.