Single-channel aliasing signal modulation mode recognition method based on residual neural network

Abstract

The invention discloses a single-channel aliasing signal modulation mode recognition method based on a residual neural network. The single-channel aliasing signal modulation mode recognition method comprises the following steps of 1, obtaining time-frequency aliasing signals of different code rates through simulation; 2, building and training a neural network; 3, acquiring actually measured aliasing signals, and only carrying out necessary signal processing on the received and transmitted signals; and 4, inputting the signals obtained in the step 3 into the residual neural network obtained in the step 2 to obtain a recognition result of the aliasing signals. According to the invention, a stacked residual error neural network is adopted, the traditional serial residual error module connection is improved, and the defects of high redundancy of information processing and unstable hierarchical transmission of a traditional residual error neural network are avoided; digital aliasing signals with different code rates are input into the network in a data augmentation form, so that the robustness and generalization ability of the network are enhanced; the method does not need any prior information, does not need to carry out additional preprocessing on the signals, and directly carries out the recognition and classification of the signals.

Description

technical field [0001] The invention relates to the technical field of signal processing, in particular to a single-channel aliasing signal modulation pattern recognition method based on a residual neural network. Background technique [0002] In non-cooperative communication systems, modulation identification of communication signals is the premise and key to process received signals. Traditional single-channel single-signal recognition techniques include wavelet transform, high-order cumulant, and instantaneous feature parameter extraction. In the existing methods based on instantaneous feature parameter extraction, the extracted parameters belong to the second-order statistics of the signal. It is less effective in low signal-to-noise ratio situations or when the channel experiences deep fading. The existing high-order cumulant can perfectly suppress the interference of noise to the signal in theory, but its effect depends greatly on the length of the received signal an...

AI Technical Summary

Problems solved by technology

Bayer and M. Using the joint maximum likelihood estimation theory, the common identification of space-time block code STBC and traditional modulation methods is realized, but these two methods depend greatly on the accuracy of prior information and have a huge amount of calculation

Images