计算机科学
遗传算法
人工神经网络
前馈
构造(python库)
光纤激光器
激光器
双折射
光学
算法
傅里叶变换
前馈神经网络
鉴定(生物学)
光纤
纤维
编码(内存)
快速傅里叶变换
信号处理
工作(物理)
迭代法
非线性系统
模式锁定
电子工程
相位恢复
序列(生物学)
进化算法
作者
Xiaoyang Shi,Changjian Lv,Qi Yan,Yifei Sun,Fanchao Meng,Zhixu Jia,Weiping Qin,Guanshi Qin
出处
期刊:Optics Express
[Optica Publishing Group]
日期:2025-11-17
卷期号:33 (24): 51307-51307
被引量:1
摘要
Bidirectional mode-locked fiber lasers have recently attracted intense research interest for dual-comb spectroscopy applications. However, achieving stable bidirectional single-pulse mode-locking typically requires careful manual adjustment of laser parameters such as birefringence and pump power, owing to the interactions and competitions between the counter-propagating pulses. In this work, we demonstrate, for the first time to the best of our knowledge, an automatically mode-locked bidirectional soliton fiber laser. A gold nanofilm based saturable absorber, featuring a high damage threshold and good long-term stability, is employed to initiate mode-locking. An iterative multi-stage genetic algorithm is developed to realize automatic mode-locking. The algorithm utilizes only the dispersive Fourier transforms of the bidirectional output pulses to construct the fitness function. Furthermore, a feedforward neural network integrated within the algorithm enables accurate identification of bidirectional single-pulse operation. This work opens new avenues for developing intelligent dual-comb light sources.
科研通智能强力驱动
Strongly Powered by AbleSci AI