光电解耦架构下自治光网络OSNR精准评估技术研究与应用展望

韩雨晴; 胡骞; 吕浙青; 朱显坤; 秦忻; 苏广楠; 龚雅栋

doi:10.11959/j.issn.1000-0801.DXKX260301

Chinese

您当前的位置：

首页 >

文章列表页 >

光电解耦架构下自治光网络OSNR精准评估技术研究与应用展望

专题：自治光网络 | 更新时间：2026-09-30

- 光电解耦架构下自治光网络OSNR精准评估技术研究与应用展望
- Research on accurate OSNR estimation and application prospects for autonomous optical networks under decoupled optical-electrical framework
- 电信科学 2026年
- 作者机构：
  
  1.中国电信股份有限公司研究院，北京 102209
  2.中电信智能网络科技有限公司，北京 102209
- 作者简介：
  
  韩雨晴（1995− ），女，硕士，中国电信股份有限公司研究院，工程师，科研人员，主要研究方向为光网络智能运维技术，光传送网数字孪生模型等。
  胡骞（1988− ），男，博士，中电信智能网络科技有限公司，高级工程师，传输网产品线副总经理，主要研究方向为开放光网络、智能光网络、光网络管控系统及相关标准制定等。
  吕浙青（1997− ），女，硕士，中国电信股份有限公司研究院，工程师，科研人员，主要研究方向为光网络智能运维技术、光网络智能化算法等。
  朱显坤（1996− ），男，硕士，中电信智能网络科技有限公司，工程师，科研人员，主要研究方向为SDN与软件系统架构设计、传输网管控技术等。
  秦忻（1995− ），女，博士，中国电信股份有限公司研究院，工程师，科研人员，主要研究方向为光网络智能运维技术、光网络通感一体技术及相关标准制定等。
  苏广楠（1987− ），男，硕士，中电信智能网络科技有限公司，高级工程师，科研人员，主要研究方向为SDN与软件系统架构设计、传输网管控技术等。
  龚雅栋（1985− ），男，硕士，中国电信股份有限公司研究院，高级工程师，传输网络研究中心副总监，主要研究方向为光传送网（OTN）与光业务单元（OSU）、高速光通信与光模块技术、智能算力网络协同等。
- 基金信息：
  
  国家重点研发计划项目(2023YFB2905400)
- DOI：10.11959/j.issn.1000-0801.DXKX260301
  中图分类号： TN915;TP302
- 收稿：2026-05-13，
  
  修回：2026-07-22，
  
  录用：2026-08-20，
- 稿件说明：
移动端阅览
韩雨晴, 胡骞, 吕浙青, 等. 光电解耦架构下自治光网络OSNR精准评估技术研究与应用展望[J/OL]. 电信科学, 2026.

HAN Yuqing, HU Qian, LV Zheqing, et al. Research on accurate OSNR estimation and application prospects for autonomous optical networks under decoupled optical-electrical framework[J/OL]. Telecommunications Science, 2026.
韩雨晴, 胡骞, 吕浙青, 等. 光电解耦架构下自治光网络OSNR精准评估技术研究与应用展望[J/OL]. 电信科学, 2026. DOI： 10.11959/j.issn.1000-0801.DXKX260301.

HAN Yuqing, HU Qian, LV Zheqing, et al. Research on accurate OSNR estimation and application prospects for autonomous optical networks under decoupled optical-electrical framework[J/OL]. Telecommunications Science, 2026. DOI： 10.11959/j.issn.1000-0801.DXKX260301.

摘要

在光电解耦架构下，光信噪比（optical signal-to-noise ratio，OSNR）评估面临精度不足、跨厂商场景适配困难和维护过度依赖人工等问题。本文围绕自治光网络中的OSNR精准评估技术开展研究，并提出配套智能管控方案。基于人工智能（artificial intelligence，AI）技术，对光层与电层参数进行解耦，结合跨域自适应对齐来缓解多厂商设备间的异构偏移；构建变换器（Transformer）与类别型梯度提升树（CatBoost）混合OSNR智能评估模型，实现OSNR高精度估计。在此基础上，将评估模型嵌入自治闭环机制，形成覆盖“数据获取—OSNR预测—异常判定—配置优化—迭代收敛—报告输出”的全流程自治体系。基于现网环境的实验结果显示，该模型OSNR评估的平均绝对误差（mean absolute error

MAE）为0.381 dB，相比传统耦合模型精度提升78.99%，验证了该方法在光电解耦架构下支撑自治光网络智能管控的可行性。

Abstract

In optical-electrical decoupled architectures

OSNR estimation methods suffer from inadequate accuracy

poor cross-vendor scenario adaptability

and heavy reliance on manual maintenance. This paper investigates precise OSNR assessment techniques for autonomous optical networks and proposes a corresponding intelligent management and control scheme. Drawing on AI technologies

we first decouple the optical-layer and electrical-layer parameters

and introduce cross-domain adaptive alignment to mitigate heterogeneous shifts across equipment from different vendors. A hybrid OSNR estimation model combining Transformer and CatBoost is then constructed to achieve high-accuracy estimation. By embedding this model into an autonomous closed-loop mechanism

we establish a full-process autonomous system

covering data acquisition

OSNR prediction

anomaly identification

configuration optimization

iterative convergence

and report generation. Experimental results from a live network environment show that the model achieves an MAE of 0.381 dB in OSNR estimation

corresponding to a 78.99% improvement in estimation accuracy over the conventional coupled model

thus verifying the feasibility of this approach for supporting intelligent management and control of autonomous optical networks under an optical-electrical decoupled architecture.

关键词

Keywords

references

FABREGA J M , et al . Experimental demonstration of a metro area network with terabit-capable sliceable bitrate variable transceiver using direct modulated VCSELs and coherent detection [C ] // Proceedings of 2022 Optical Fiber Communications Conference and Exhibition (OFC) . Piscataway : IEEE Press , 2022 : 1 - 3 .

JIANG X T , QIN X , et al . Grey-box digital twin model for QoT estimation in field-deployed networks [C ] // Proceedings of 2026 Optical Fiber Communication Conference (OFC) . Piscataway : IEEE Press , 2026 : 1 - 3 .

BORRACCINI G , et al . Experimental demonstration of partially disaggregated optical network control using the physical layer digital twin [J ] . IEEE Transactions on Network and Service Management , 2023 , 20 ( 3 ): 2343 - 2355 .

D'AMICO A , et al . Enhancing lightpath QoT computation with machine learning in partially disaggregated optical networks [J ] . IEEE Open Journal of the Communications Society , 2021 , 2 : 564 - 574 .

DONG Z , et al . Optical performance monitoring: A review of current and future technologies [J ] . Journal of Lightwave Technology , 2016 , 34 ( 2 ): 525 - 543 .

MORETTE N , et al . Machine learning enhancement of a digital twin for wavelength division multiplexing network performance prediction leveraging quality of transmission parameter refinement [J ] . Journal of Optical Communications and Networking , 2023 , 15 ( 6 ): 333 - 343 .

PAOLUCCI F , et al . Peer-to-peer disaggregated telemetry for autonomic machine-learning-driven transceiver operation [J ] . Journal of Optical Communications and Networking , 2022 , 14 ( 8 ): 606 - 620 .

LI J J , et al . Reservoir computing for few-mode fiber channel OSNR monitoring in deep learning frameworks [J ] . IEEE Photonics Journal , 2025 , 17 ( 2 ): 1 - 7 .

AHSAN M M , et al . Effect of data scaling methods on machine learning algorithms and model performance [J ] . Technologies , 2021 , 9 ( 3 ): 52 .

杨寒雨 , 赵晓永 , 王磊 . 数据归一化方法综述 [J ] . 计算机工程与应用 , 2023 , 59 ( 3 ).

YANG H Y , ZHAO X Y , WANG L . Review of data normalization methods [J ] . Computer Engineering and Applications , 2023 , 59 ( 3 ).

ALHADIDI T I , et al . Enhancing pavement crack classification with bidirectional cascaded neural networks [C ] // Proceedings of 2025 IEEE International Conference on Advanced Systems and Emergent Technologies (IC_ASET) . Piscataway : IEEE Press , 2025 : 1 - 6 .

VASWANI A , et al . Attention is all you need [C ] // Proceedings of 2017 Neural Information Processing Systems (NIPS) . Long Beach : Curran Associates Inc. , 2017 .

GRETTON A , et al . A kernel two-sample test [J ] . Journal of Machine Learning Research , 2012 , 13 : 723 - 773 .

PROKHORENKOVA L , et al . CatBoost: unbiased boosting with categorical features [C ] // Proceedings of the 32nd International Conference on Neural Information Processing Systems (NeurIPS) . Piscataway : IEEE Press , 2018 : 6639 - 6649 .

CHINA COMMUNICATIONS STANDARDS ASSOCIATION . General technical requirements for optical wavelength division multiplexing systems : YD/T 3391-2018 [S ] . 2018 .

CHINA COMMUNICATIONS STANDARDS ASSOCIATION . Technical specification for optical safe procedure for Wavelength Division Multiplexing (WDM)system : YD/T 1259-2003 [S ] . 2003 .

叶胤 , 汪令全 . 骨干光网络工程系统性能指标分析 [J ] . 邮电设计技术 , 2022 ( 11 ): 65 - 69 .

YE Y , WANG L Q . Analysis of system performance indexes in backbone optical network engineering [J ] . Designing Techniques of Posts and Telecommunications , 2022 ( 11 ): 65 - 69 .

WANG Z , et al . Integrated sensing and communication (ISAC) in optical fiber networks: Current research and challenges [C ] // Proceedings of 2025 International Conference on Advanced Information and Communication Technology (ICAIT) . Piscataway : IEEE Press , 2025 : 1 - 5 .

BAO T , et al . Joint optimization of DNN model partitioning and slice delivery for distributed edge-cloud inference over optical networks [J ] . Journal of Optical Communications and Networking , 2025 , 17 ( 11 ): 1047 - 1058 .

浏览量

下载量

CSCD

文章被引用时，请邮件提醒。

提交

工具集

关联资源

电信DPI技术与新媒体数据融合的受众态势分析模型研究

AI赋能工单与投诉自动化分类的研究与实践

VisionTS 时序预测在电信智能决策辅助系统中的应用与实现

可信数据空间在医疗行业应用的研究与探索

长文本动态感知分块与语义嵌入协同优化算法