Spectrum sensing method and system based on CGAN data enhancement

Abstract

The invention provides a spectrum sensing method and system based on CGAN data enhancement, and the method comprises the steps: collecting channel state data, and obtaining a training data set; through short-time Fourier transform, converting the training data into a spectrogram, and generating a training set; establishing a neural network model based on the CGAN, and training the neural network model based on the CGAN through the training set; adding a data sample generated by the trained CGAN-based neural network model into the training set to generate an enhanced training data set; and obtaining a pre-trained convolutional neural network, and inputting the enhanced training data set into the convolutional neural network for classification to obtain a classification result, wherein the classification result is that a master user of a channel is in a silent state or in an active state. According to the invention, after the CGAN network is adopted to enhance the data, the AlexNet convolutional neural network is introduced for classification, so that spectrum sensing can be effectively realized.

Description

technical field [0001] The present invention relates to cognitive radio technology, in particular, to a spectrum sensing method and system based on CGAN data enhancement. Background technique [0002] With the development of emerging wireless communication technologies such as 5G and the Internet of Things, the demand for wireless spectrum resources is increasing. The Federal Communications Commission (FCC) of the United States once researched and pointed out that a large number of allocated spectrum resources are largely idle in both time and space. The main reasons for the shortage of spectrum resources are as follows: 1. The static allocation policy of spectrum resources. 2. Inefficient use of certain licensed frequency bands. Cognitive Radio Networks (CRNs) are considered to be an effective way to achieve dynamic allocation of spectrum resources and improve spectrum utilization. [0003] In CRNs, there are two types of users, one is the primary user (PU) and the other...

AI Technical Summary

Problems solved by technology

2. Inefficient use of certain licensed frequency bands

However, these deep learning models based on supervised learning require a large number of training data samples for training. In the actual complex channel environment conditions, it may not be possible to obtain enough data with labels, and to collect sufficient data. The cost is usually very expensive, which leads to poor generalization ability of the model trained from a small number of samples

Images