Generative Adversarial Network-based Channel Estimation in High-Speed Mobile Scenarios
Danyang Zhang, Junhui Zhao, Lihua Yang, Yiwen Nie, Xiangcheng Lin
Abstract
To reduce the complexity of the channel estimation method and improve the estimation performance in the high-speed mobile scenario, a generative adversarial network (GAN) based channel estimation method is proposed to ensure the stability of the high-speed railway communication system with the fast time-varying and non-stationary channel. Firstly, we use the GAN's discriminator to learn and extract channel time-varying features. And the standard wireless high-speed mobile channel data is used for offline training of the network. After that, the GAN's generator is used to extract the features of high-dimensional time-varying channel to generate and restore channel information close to reality. The simulation results show that the proposed channel estimation method can effectively extract the characteristic and distribution of fast time-varying channel and predict the channel response.