OSNR monitoring based on a low-bandwidth coherent receiver and LSTM classifier

• DOI: 10.1364/oe.412079
• Publication Date: 2020-12-22T12:30:08Z
• Authors: Huangbin Ye; Hao Jiang; Guowei Liang; Qianxin Zhan; Shuqiang Huang; Dawei Wang; Hongwei Di; Zhaohui Li
• Keywords: Robustness

Optical signal-to-noise ratio (OSNR) monitoring is one of the core tasks advanced optical performance (OPM) technology, which plays an essential role in future intelligent communication networks. In contrast to many regression-based methods, we convert continuous OSNR into a classification problem by restricting outputs neural network-based classifier discrete intervals. We also use low-bandwidth coherent receiver for obtaining time domain samples and long short-term memory (LSTM) network as chromatic dispersion-resistant classifier. The proposed scheme cost efficient compatible with our previously multi-purpose OPM platform. Both simulation experimental verification show that technique achieves high accuracy robustness low computational complexity.