A New Method for OFDM Signal Modulation Recognition Based on Optimized Convolutional Neural Network

Due to the rapid development of communication technology, the environment in which electromagnetic signals are located is complicated and intricate. For the problem of OFDM signal modulation recognition in a complex electromagnetic environment, this paper proposes a new method based on deep learning model. Firstly, the structure of convolutional neural network is optimized and the model parameters are designed to be more suitable for the actual environment and the structure of the convolutional neural network is optimized, and then the convolutional neural network and Transformer network model are integrated to enhance the recognition effects, thus making the new method more robust and noise-resistant. The simulation demonstrate that the proposed method has better recognition performance in low signal-to-noise ratio situations, and the ability to improve the performance of OFDM signal modulation recognition.