DFENet: A Diverse Feature Extraction Neural Network for Improving Automatic Modulation Classification Accuracy in Wireless Communication Systems

This is a validated dataset comprising 24 digital and analog modulation types, generated with synthetic simulated channel effects to emulate realistic wireless transmission conditions. These channel impairments include AWGN, fading, frequency offset, and phase offset. Each signal sample is represented as a complex IQ sequence of length 1024, consisting of in-phase and quadrature components, and is labeled with SNR values ranging from −20 dB to 30 dB in 2 dB steps. Owing to its large scale and balanced SNR distribution, the dataset has become a benchmark for evaluating the robustness and generalization capability of deep learning–based AMC methods. If this dataset is used, please cite the paper "<i>DFENet: A Diverse Feature Extraction Neural Network for Improving Automatic Modulation Classification Accuracy in Wireless Communication Systems".</i>