Abstract
An artificial neural network (ANN) based maximum-likelihood (ML) symbol detection is proposed for nonlinearity compensation in PAM-4. Its optimized architecture realizes 0.7-dB better sensitivity and 20-times lower complexity than the conventional ANN-based ML sequence detection.
© 2018 The Author(s)
PDF ArticleMore Like This
Xiaoxiao Dai, Xiang Li, and Ming Luo
W1I.4 Conference on Lasers and Electro-Optics/Pacific Rim (CLEO/PR) 2018
Ahmed Galib Reza and June-Koo Kevin Rhee
Su3C.5 Asia Communications and Photonics Conference (ACP) 2018
Ming Luo, Fan Gao, Xiang Li, Zhixue He, and Songnian Fu
M2F.2 Optical Fiber Communication Conference (OFC) 2018