Electromagnetic modulation signal denoising method and system based on deep learning

Abstract

The invention discloses an electromagnetic modulation signal denoising method and system based on deep learning. The method comprises the steps of S1, making a target signal and noise signal data set; S2, using a filtering and denoising method and a signal enhancement method to expand and optimize a data set; S3, respectively defining model structures and loss functions of the generator and a discriminator, and training the model until the model is stable; and S4, outputting a denoising result. A generator loss function mentioned in the method combines and uses discriminator output loss, minimum absolute value deviation and continuity difference, and particularly considers the continuity characteristic of a generated de-noised signal. The invention further comprises an electromagnetic modulation signal denoising system based on deep learning. The system is composed of a data processing module, a training module and an output module which are connected in sequence. A filtering and denoising method not based on learning and a denoising method based on learning are combined, signal characteristics can be adaptively learned, signal denoising is achieved, and the invention has good universality in signal denoising.

Description

technical field [0001] The invention relates to the field of unsupervised learning of electromagnetic signals, in particular to a method and system for denoising electromagnetic modulation signals based on deep learning. Background technique [0002] Electromagnetic signals refer to electromagnetic waves that propagate in free space, such as radio signals. In signal processing, the presence of noise is a common problem with any signal type. Neural network is a learning-based approach to signal denoising that does not require precise modeling of signal and noise, nor optimal parameter tuning. This approach is especially popular in the field of image-based learning. Similarly, in audio and speech processing, deep neural networks have also achieved good results. In contrast, learning-based methods have rather limited effects on low-dimensional signals such as modulated signals. For low-dimensional signals, previous denoising work mainly uses non-learning methods such as fil...

AI Technical Summary

Problems solved by technology

However, this method does not study the modulated signal, and does not consider combining the non-learning-based filter denoising method with the learning-based method to improve the signal denoising effect

Images