Automatic modulation identification method based on high signal-to-noise ratio pre-training

• Main Authors: AN JIANPING, DING HAICHUAN, LI YIYANG, DING XUHUI, MA YING
• Format: Patent
• Language: Chinese; English
• Published: 12.03.2024
• Subjects: CALCULATING; COMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS; COMPUTING; COUNTING; ELECTRIC DIGITAL DATA PROCESSING; PHYSICS

The invention discloses an automatic modulation identification method based on high signal-to-noise ratio pre-training, and belongs to the technical field of communication signal identification. In order to solve the problems that in an existing deep learning automatic modulation recognition method, a low-signal-to-noise-ratio signal affects network feature extraction performance and recognition precision is reduced, high-signal-to-noise-ratio data is used for pre-training, low-signal-to-noise-ratio data is added for fine tuning, the influence of the low-signal-to-noise-ratio data on neural network recognition precision in the training process is reduced, and the recognition precision of a neural network is improved. Therefore, the recognition accuracy of the automatic modulation recognition technology is improved. Through task target division, sub-network training and a cascaded network structure, decomposition of a multi-classification automatic modulation identification problem is realized, so that the aut