基于注意力机制的OSNR计算方法  

OSNR calculation method based on attention mechanism

作　　者：李文超 匡立伟 戴非 LI Wenchao;KUANG Liwei;DAI Fei(Wuhan Research Institute of Posts and Telecommunication,Wuhan 430070,China;FiberHome Telecommunication Technologies Co.Ltd.,Wuhan 430070,China;School of Computer Science,Wuhan University,Wuhan 430070,China)

机构地区：[1]武汉邮电科学研究院,武汉430070 [2]烽火通信科技股份有限公司,武汉430070 [3]武汉大学计算机学院,武汉430070

出　　处：《光通信技术》2020年第12期52-55,共4页Optical Communication Technology

摘　　要：光信噪比(OSNR)计算是光网络健康评估和性能预测的关键指标,但是传统计算方法在精确度与鲁棒性方面面临很大挑战。在循环神经网络的基础上,提出基于注意力机制的OSNR计算方法,该方法通过直接捕捉时序输入与OSNR之间的逻辑关系,提高OSNR的计算精度。针对不同入纤功率、传输距离和解调方式,设计了多种类型的实验,结果表明提出的方法能够将OSNR计算误差降低至0.36 dB以下。The optical signal to noise ratio(OSNR)is one of the key indexes for health assessment and performance prediction of optical network.However,traditional methods face great challenges for computing accusation and system robustness.Based on the recurrent neural network and attention mechanism,this paper presents an OSNR calculation method based on attention mechanism.This method directly captures the logical relationship between the timing input and the OSNR,and improves the calculation accuracy of OSNR.According to different fiber input power,transmission distance and demodulation methods,a variety of experiments are designed.The results show that the proposed method can reduce the calculation error of OSNR to less than 0.360 dB.

关 键 词：光网络 神经网络 注意力机制 光信噪比 性能预测 

分 类 号：TN929.1[电子电信—通信与信息系统]