基于眼图重构和人工神经网络的光性能监测  

Optical performance monitoring based on reconstructed eye diagrams and artificial neural networks

作　　者：赖俊森[1] 杨爱英[1] 孙雨南[1] 

机构地区：[1]北京理工大学光电学院,北京100081

出　　处：《光电子．激光》2011年第9期1342-1347,共6页Journal of Optoelectronics·Laser

基　　金：国家自然科学基金资助项目(60978007)

摘　　要：提出了一种基于异步降频光采样眼图重构和人工神经网络(ANN)的光性能监测(OPM)新方法。首先对被监测光信号进行异步降频光采样,通过软件同步算法进行眼图重构;然后提取重构眼图的特征参数对ANN进行训练;最后以ANN的预测输出对光信号的损伤进行监测。构建10 Gb/s NRZ-OOK4、0 Gb/s RZ-OOK和40 Gb/s RZ-DPSK仿真实验系统,进行光信噪比(OSNR)和色散(CD)参数监测。结果表明,本文方法进行OPM具有较高的精度,ANN预测输出与测试数据的相关系数大于0.98,损伤监测的平均误差小于5%。A novel optical performance monitoring（OPM） method based on asynchronous optical-sampling,eye diagram reconstruction and artificial neural network（ANN） is presented.Firstly,the monitored optical signal is optically sampled in asynchronous way,and the eye diagrams are reconstructed by software-synchronized algorithm.Secondly,the features of reconstructed eye diagrams are extracted to train the artificial neural network.Finally,the outputs of the trained neural network are used to monitor optical signal impairments.Simulations of optical signal noise ratio（OSNR） and chromatic dispersion（CD） monitored in 10 NRZ-OOK,40 Gbit/s RZ-OOK and 40 Gbit/s RZ-DPSK systems are presented.The monitoring results show that the accuracy of this proposed OPM method is higher,the correlation coefficient between neural network output and test data is greater than 0.98,and the impairment monitoring average error is less than 5%.

关 键 词：重构眼图 人工神经网络(ANN) 光性能监测(OPM) 光信噪比(OSNR) 色散(CD) 

分 类 号：TN929.11[电子电信—通信与信息系统]