Research on Automatic Modulation Classification Technology under α Stable Distribution Noise

The Automatic modulation classification is the base and significant part of spectrum utilization and the corresponding design of receiver. Therefore, the influence of automatic modulation classification complex electromagnetic environment is very important. In this paper, we consider the diversity of signal noise in complex electromagnetic environment. So we construct the directory in the same modulation signal containing different signals under different Mixed signal-to-noise (MSNR), use the deep learning model, and finally the classification will be obtained. This method only needs to train one model to recognize modulation classification, which avoids the redundancy of model training which trains in different SNR, and the robustness of the network is improved.