Signal Modulation Recognition of Satellite Internet Based on Dual Input Convolutional Neural Network
Yiyu Pan, Feng Gao · 2024
In order to identify the main modulation modes used in satellite Internet communication systems, a communication signal modulation recognition algorithm based on double-ended convolutional neural network is proposed. The time domain waveform of the signal is transformed into the high-power spectrum and the time-frequency image, and further converted into gray images as the shallow feature expressions of the signal. The modulation recognition problem is converted into an image recognition problem, and a modulation recognition model based on dual input convolutional neural network is designed. Through the training of the network, the shallow features are extracted and mapped deeply, and the modulation recognition of the target signal is finally completed. The simulation results show that the performance of the proposed algorithm is significantly improved compared with the single feature input modulation recognition algorithm, and it has good recognition performance for satellite Internet signals