Abstract
Two deep learning-based differentiable models based on GAN and BiLSTM are proposed for FSO channel modeling. Through contrastive analysis, it shows that GAN is more suitable for learning FSO channel distribution with KL divergence<0.55.
© 2021 The Author(s)
PDF ArticleMore Like This
Guo Wenjing, Shi Ziyuan, Zhan Yueying, and Yang Lei
S4B.6 Optoelectronics and Communications Conference (OECC) 2021
Song Song, Yejun Liu, Tianming Xu, Shasha Liao, and Lei Guo
Th1A.36 Optical Fiber Communication Conference (OFC) 2021
Chenda Lu, Qinghua Tian, Xiangjun Xin, Lei Zhu, Qi Zhang, Haipeng Yao, Huan Chang, and Ran Gao
Th1A.16 Optical Fiber Communication Conference (OFC) 2021