Abstract
We propose a FPGA-based artificial neural network (ANN) architecture with flexible degree of parallelism (DOP) for nonlinear signal-to-noise ratio (SNR) estimation. Considering the trade-off between resource usage and computational speed, the architecture shows flexibility and efficiency of resources allocation in the FPGA design.
© 2021 The Author(s)
PDF ArticleMore Like This
Aazar S. Kashi, John C. Cartledge, and Wai-Yip Chan
M5F.3 Optical Fiber Communication Conference (OFC) 2021
Aazar Saadaat Kashi, Qunbi Zhuge, John Cartledge, Andrzej Borowiec, Douglas Charlton, Charles Laperle, and Maurice O'Sullivan
Su1B.6 Asia Communications and Photonics Conference (ACP) 2017
Feng Wang, Shencheng Ni, Shuying Han, Shanhong You, Ming Luo, and Zhixue He
M4A.246 Asia Communications and Photonics Conference (ACP) 2020