Mitigating ambiguity by deep-learning-based modal decomposition method

Fan, Xinyan; Ren, Fang; Xie, Yulai; Niu, Jingjing; Zhang, Jingyu; Wang, Jianping

doi:10.1016/j.optcom.2020.125845

Cited by 20 publications

(10 citation statements)

References 29 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…In 2020, Xiaojie Fan et al handled the phase ambiguity of modal coefficients using two labels: near-field and far-field images [121]. A combined loss to train a convolutional neural network considers the reconstruction loss of near-field and far-field images together.…”

Section: Mode Decompositionmentioning

confidence: 99%

Fiber laser development enabled by machine learning: review and prospect

Jiang

et al. 2022

PhotoniX

View full text Add to dashboard Cite

In recent years, machine learning, especially various deep neural networks, as an emerging technique for data analysis and processing, has brought novel insights into the development of fiber lasers, in particular complex, dynamical, or disturbance-sensitive fiber laser systems. This paper highlights recent attractive research that adopted machine learning in the fiber laser field, including design and manipulation for on-demand laser output, prediction and control of nonlinear effects, reconstruction and evaluation of laser properties, as well as robust control for lasers and laser systems. We also comment on the challenges and potential future development.

show abstract

Section: Mode Decompositionmentioning

confidence: 99%

Fiber laser development enabled by machine learning: review and prospect

Jiang

et al. 2022

PhotoniX

View full text Add to dashboard Cite

show abstract

“…Deep learning concepts show a lot of potential to solve mode analysis problems numerically and are recently frequently researched. Some aspects of uncertainty might be mitigated with a proper choice of neural network [36,37]. Although, to our knowledge, these techniques were only applied to fewmode fibers, we believe they are equally beneficial for the analysis of highly multi-mode fiber beams, in particular, if computation time and the minimization of inaccuracies in the determination of mode coefficients is considered.…”

Section: Challenges Of Numerical Mode Analysis Techniquesmentioning

confidence: 99%

Focal field analysis of highly multi-mode fiber beams based on modal decomposition

et al. 2020

View full text Add to dashboard Cite

In this work, a numerical modal decomposition approach is applied to model the optical field of laser light after propagating through a highly multi-mode fiber. The algorithm for the decomposition is based on the reconstruction of measured intensity profiles along the laser beam caustic with consideration of intermodal degrees of coherence derived from spectral analysis. To enhance the accuracy of the model, different approaches and strategies are applied and discussed. The presented decomposition into a set of LP modes enables both the wave-optical simulation of radiation transport by highly multi-mode fibers and, additionally, the analysis of free-space propagation with arbitrarily modified complex amplitude distributions.

show abstract

“…27,28 Nevertheless, the coupling problem of the DQN target network and the Q network will cause convergence issues when dealing with specific problems, which is particularly fatal in embedded systems. 29 The computational burden and the demanding training data set also make it full of challenges. 30…”

Section: Introductionmentioning

confidence: 99%

Pressure control based on reinforcement learning strategy of the pneumatic relays for an electric-pneumatic braking system

Shan

et al. 2022

Proceedings of the Institution of Mechanical Engineers, Part D:

View full text Add to dashboard Cite

In the electric-pneumatic braking system (EPBS), fast and accurate brake pressure regulation is critical to vehicle braking safety and is the basis for active safety functions. However, the lack of signal feedback, limited actuator response accuracy, and extremely strong model stiffness and nonlinearity pose problems for high-precision brake pressure regulation. To solve these problems, this article proposes a Q-learning-based control algorithm to regulate actuator instructions. First, the nonlinearity of the system and complicated actuator operating process is settled by a simplified mathematical model. Then, a decoupled pressure observing method is proposed to solve the observation problem caused by the coupling of mechanical and fluid motion. Finally, this paper proposes the idea of using an optimization method to solve the control problem caused by the response speed of the actuator, a Q-learning algorithm is used to settle the solenoid switching action to minimize response time and steady-state error. Both simulation and practical experiments are conducted to demonstrate the dependability of the model and effectiveness of the control method.

show abstract

Mitigating ambiguity by deep-learning-based modal decomposition method

Cited by 20 publications

References 29 publications

Fiber laser development enabled by machine learning: review and prospect

Fiber laser development enabled by machine learning: review and prospect

Focal field analysis of highly multi-mode fiber beams based on modal decomposition

Pressure control based on reinforcement learning strategy of the pneumatic relays for an electric-pneumatic braking system

Contact Info

Product

Resources

About