3D convolutional neural networks based automatic modulation classification in the presence of channel noise

Automatic modulation classification is a task that essentially required in many intelligent communication systems such as fibre-optic, next-generation 5G or 6G systems, cognitive radio well multimedia internet-of-things networks etc. Deep learning (DL) representation method takes raw data and finds representations for different tasks detection. DL techniques like Convolutional Neural Networks (CNNs) have strong potential to process analyse large chunks of data. In this work, we considered the problem multiclass (eight classes) modulated signals, which are, Binary Phase Shift Keying, Quadrature 16 64 Amplitude Modulation corrupted by Additive White Gaussian Noise, Rician Rayleigh fading channels using 3D-CNN architectures both frequency spatial domains while deploying three approaches augmentation, random zoomed in/out, shift weak blurring augmentation with cross-validation (CV) based hyperparameter selection statistical approach. Simulation results testify performance 10-fold CV without domain be best worst performing happens found better than domain.