6G automatic modulation classification using deep learning models in the presence of channel noise, CFO, and PN

An efficient and remarkable automatic modulation classification (AMC) technique is essential with the advent of sixth-generation (6G) communication systems. Using the pre-trained convolutional neural network (CNN), a deep learning (DL) approach to classify eight types of digital modulated signals. National Instrument LabVIEW NXG is used to build the modulation transceivers at 100 GHz, a 6G carrier frequency. The dataset was collected in a complicated environment, including carrier frequency offset (CFO), phase noise (PN), and distinct signal-to-noise ratios (SNR). Through experimental simulation, an improvement in the classification accuracies was achieved. In particular, the outstanding accuracy rates achieved are 98.68% and 96.05% using ResNet18 and ResNet101, respectively. Furthermore, these models can classify the modulated signals at lower SNRs. These innovative models are suitable and effective to utilize for 6G wireless communication networks.