Signal modulation identification method and system combining deep neural network and expert prior characteristics

Abstract

The invention belongs to the field of radio communication, and particularly relates to a signal modulation identification method and system combining a deep neural network and expert prior features, and the method comprises the steps: constructing a signal modulation identification network model, and carrying out the model training through the signal sample data of a known signal modulation type, the model comprises a first-level neural network used for carrying out inter-class classification according to different feature parameters, and a second-level neural network connected with the first-level neural network and used for carrying out intra-class classification by extracting internal signal features of corresponding subclasses according to inter-class classification results of different feature parameters. The number of the secondary neural networks is consistent with the classification number between feature parameter classes; and selecting a plurality of target signal characteristic parameters for signal modulation identification according to the dependence degree and the discrimination degree on the signal prior information, and identifying the modulation category of the target signal by using the trained signal modulation identification network model. According to the invention, the recognition accuracy is improved, and the signal recognition effect is ensured under the influence of multipath fading and frequency offset.

Description

technical field [0001] The invention belongs to the field of radio communication, and in particular relates to a signal modulation recognition method and system combining deep neural network and expert prior features. Background technique [0002] Signal modulation identification is an essential part of signal interception analysis, and has always been a research hotspot in the field of radio communication. As the electromagnetic spectrum environment becomes increasingly complex, modulation methods emerge in endlessly. How to quickly and accurately identify the modulation methods of different signals plays a key role in whether subsequent demodulation and analysis can be performed. At present, the traditional automatic modulation recognition technology is mainly the recognition method based on the prior characteristics of experts. The method is mainly divided into two steps: feature extraction and pattern recognition. The features that can represent the modulation mode are ...

AI Technical Summary

Problems solved by technology

Automatic modulation recognition based on deep learning is one of the earlier and more mature fields of deep learning in radio communication, and it has been deployed in practical applications, such as: using Convolutional Neural Networks (CNN) to modulate signals Classification and recognition, and achieved good results, but when the input data is time-series data with memory, CNN's performance is not good

Images