-
Aiming at the challenging problem of analyzing the strong nonlinear coupling effect between four-wave mixing and stimulated Raman scattering in single-mode optical fibers, this paper introduces a novel multi-scale physically constrained network, designated as MSPC-Net, which effectively integrates fundamental physical mechanisms with advanced neural network techniques. The proposed model incorporates the frequency domain residual derived from the nonlinear Schrödinger equation directly into the network optimization procedure as a differentiable physical constraint term. This strategic inclusion ensures that the learning process remains consistent with the underlying physical principles governing light propagation in optical fibers. Furthermore, the model architecture employs a multiscale dilated convolution module specifically designed to capture and fuse features across different granularities, including fine local spectral details, intermediaterange broadening effects, and long-range attenuation trends. This multi-scale approach enables the simultaneous and high-precision inversion of both separated spectral components and critical physical parameters.
Experimental evaluations were conducted using single-mode quartz fibers with lengths of 250 meters and 500 meters. The results demonstrate that the Stokes spectra reconstructed by MSPC-Net achieve remarkably low root mean square errors, measuring only 0.014 and 0.0173 for the two fiber lengths respectively. This performance represents a reduction of more than sixty-eight percent compared to conventional convolutional neural networks. Additionally, the average absolute errors for frequency offset prediction are as low as 0.03 nanometer and 0.04 nanometer, corresponding to an accuracy improvement of approximately ninety percent relative to existing state-of-the-art methods. Under noisy conditions with a signal-to-noise ratio of 6 decibels, the model maintains an exceptional detection accuracy of up to 95.3 percent for identifying FWM sub-peak information, while keeping the pseudo-peak rate below 4.7 percent.
Benefiting from the strong guidance provided by embedded physical constraints and its lightweight structural design, the proposed model exhibits only a 9.8 percent increase in root mean square error even under challenging noise conditions with a signal-to-noise ratio of 15 decibels. Moreover, MSPC-Net demonstrates satisfactory real-time processing capabilities, making it suitable for deployment on embedded devices. This practical efficiency positions the model as a promising solution for optimizing high-power optical communication systems and advancing distributed optical fiber sensing applications. By successfully combining rigorous physical laws with multi-scale feature extraction, this research provides an effective approach to resolving the analytical difficulties associated with complex nonlinear effects in long-distance optical fibers, while significantly enhancing both the theoretical consistency and noise robustness of the prediction outcomes.-
Keywords:
- Nonlinear optics /
- Physically constrained neural networks /
- multi-scale feature extraction /
- spectral separation
-
[1] Yao T P, Fan C C, Hao X L, Li Y, Huang S W, Zhang H W, Xu J M, Ye J, Leng J Y, Zhou P 2024 Chin. J. Lasers. 51 182 (in Chinese) [姚天甫, 范晨晨, 郝修路, 李阳, 黄善旻, 张汉伟, 许将明, 叶俊, 冷进勇, 周朴 2024 中国激光51 182]
[2] Zhang F, Li J, Li L L, Cao K Y, Xue X H, Zhang M J 2025 IRLAE. 54 289 (in Chinese) [张帆, 李健, 李璐磊, 曹康怡, 薛晓辉, 张明江 2025 红外与激光工程 54 289]
[3] Zhang P, Tian C L, Qiao Y, Lyu D D 2018 LOP. 55 351 (in Chinese) [张鹏,田春林,乔勇,吕栋栋 2018 激光与光电子学进展 55 351]
[4] Zhang P, Tian C L 2016 AOP. 36 244 (in Chinese) [张鹏,田春林 2016光学学报 36 244]
[5] Mao X R, Kou Z F, Zhang J H 2017 LOP. 54 95(in Chinese) [毛昕蓉,寇召飞,张建华 2017激光与光电子学进展 54 95]
[6] Zheng Y, Ni Q L, Zhang L, Liu X X, Wang J L, Wang X F 2021 Chin. J. Lasers. 48 32(in Chinese) [郑也,倪庆乐,张琳,刘小溪,王军龙,王学锋 2021中国激光 48 32]
[7] Sui H, Zhu H N, Jia H Y, Ou M Y, Li Q, Luo B, Zou Xi H 2023 Chin. J. Lasers. 50 1101011(in Chinese) [隋皓,朱宏娜,贾焕玉,欧洺余,李祺,罗斌,邹喜华 2023 中国激光 50 70]
[8] Zhang L L, Li X R, Liu J H, Fang Q Z 2025 Chin. J. Lasers. 52 214 (in Chinese) [张丽丽,栗相如,刘佳辉,房启志2025中国激光 52 214]
[9] Jin Z C, Jia K, Li H X, Xu C Y, Wang W R, Zhou J 2025 CTAD. (in Chinese) [金治成,贾可,李涵鑫,许昌源,王文润,周记 2025 计算机技术与发展]
[10] Wang Y, Wang Z Y 2024 J. Lanzhou Univ. Technol. 50 87 (in Chinese) [王燕,王振宇 2024 兰州理工大学学报 50 87]
[11] Ding Y X, Jia M, Gu S Y, Qiu J X, Chen G H 2024 Chin. J. Lasers. 51 1901011 (in Chinese) [丁亚茜,贾明,顾劭忆,邱佳欣,陈光辉 2024 中国激光 51 201]
[12] Cai B T, Huang W T, Xiao L M, Chen X B 2025 Chin. J. Lasers. 52 164 (in Chinese) [蔡冰涛,黄文涛,肖力敏,陈小宝 2025 中国激光 52 164]
[13] Yao G J, Li J C, Liu H Z, Ma C J, Hong H, Liu K H 2025 Chin. J. Lasers. 52 106(in Chinese) [姚光杰,李家成,刘华展,马超杰,洪浩,刘开辉 2025 中国激光 52 106]
[14] Shang X J, Li S L, Ma B, Chen Y, He X W, Ni H Q, Zhi C 2021 Acta Phys. Sin. 70 087801 (in Chinese) [尚向军,李叔伦,马奔,陈瑶,何小武,倪海桥,智川2021物理学报70 087801]
[15] Hou Y, Xiang S Y, Zou T, Huang Z Q, Shi S X, Guo X X, Zhang Y H, Zheng L, Hao Y 2025 Acta Phys. Sin. (in Chinese) [侯悦, 项水英, 邹涛, 黄志权, 石尚轩, 郭星星, 张雅慧, 郑凌, 郝跃 2025 物理学报]
[16] Wu J, Cui C F, Ou Y T, Tang C 2023 Acta Phys. Sin. 72 047201 (in Chinese) [伍静, 崔春凤, 欧阳滔, 唐超 2023 物理学报72 047201]
[17] Liu Y K, Hou Y L, Yang Y L, Hou L M, Li Y H, Lin J, Chen X F 2025 Acta Phys. Sin. (in Chinese) [刘圆凯, 侯云龙, 杨宜霖, 侯刘敏, 李渊华, 林佳, 陈险峰. 基于超纠缠的三用户全连接量子网络 2025 物理学报]
[18] Qin J 2023 Acta Phys. Sin. 72 050302 (in Chinese) [覃俭 2023 物理学报 72 050302]
[19] Wang X, Zhou Y S, Zhang X G, Chen X H 2025 Acta Phys. Sin. 74 084202(in Chinese) [王翔, 周义深, 张轩阁, 陈希浩 2025 物理学报 74 084202]
[20] Wei Y X, Yang C G, Wei A M, Zhang G F, Qin C B, Chen R Y, Hu J Y, Xiao L T, Jia S T 2025 Acta Phys. Sin. 74 064208 (in Chinese) [卫祎昕, 杨昌钢, 卫阿敏, 张国峰, 秦成兵, 陈瑞云, 胡建勇, 肖连团, 贾锁堂 2025 物理学报 74 064208]
Metrics
- Abstract views: 126
- PDF Downloads: 1
- Cited By: 0