GAN-based Wireless Channel Recognition Enhancement in Aerospace Communication System

Abstract

Abstract In the field of communications, wireless channel identification is of great significance for spectrum identification and spectrum resource scheduling, and is an indispensable link in cognitive radio technology. However, the poorness of the aerospace data set will affect the recognition accuracy. This paper studies the application of neural network in aerospace communication wireless channel recognition scenarios. Then, we propose a sample set expansion method based on GAN (Generative Adversarial Networks) to enhance the cognitive ability of neural network. Finally, we compare the accuracy of the channel recognition model before and after the data set expansion. The results show that in the case of small sample data sets, the use of GAN-based data expansion method helps to improve the accuracy of the channel recognition model.