Wuli Hu
Research Subjects
Wuli Hu has a total of 4 co-authors. Here are the top co-authors by number of shared publications:
- Bin Zhang
College of Electronics and Information Engineering Sichuan University Chengdu Sichuan 610064 China
1 shared publication - Tianzhe Jia1 shared publication
- Yonghao Chen
School of Science, Shenyang Ligong University 1 , Shenyang 110159, People’s Republic of China
1 shared publication Zheqiang Zhong
1 shared publication
Latest Publications
- Open Access
- 22 October 2026
- Published by Elsevier in SSRN Electronic Journal
Physics-Constrained Deep Learning for Cross-Wavelength Wavefront Estimation
In laser beam propagation through atmospheric turbulence, the wavelength mismatch between beacon and signal beams fundamentally limits adaptive optics (AO), as both refractive-index accumulation and Fresnel diffraction jointly produce phase discrepancies that cannot be captured by any global scaling law. In high-power regimes, absorption-induced thermal blooming further breaks linear scaling. To address this ill-posed inverse problem, we propose Physics-Aware Cross-Wavelength Network (PAX-Net), a physics-informed network that maps dual-plane beacon intensities at one wavelength to Zernike coefficients at another, embedding differentiable Fresnel transfer-function propagation and phase smoothness constraints as physical priors during training. Simulations demonstrate accurate estimation in both turbulence‐dominated and thermal‐blooming conditions, where conventional scaling fails, confirming that physics‐constrained deep learning offers a viable high‐precision cross‐wavelength wavefront estimation for adaptive optics.
- 22 October 2026
- Published by Elsevier in SSRN Electronic Journal
Open AccessPhysics-Constrained Deep Learning for Cross-Wavelength Wavefront Estimation
In laser beam propagation through atmospheric turbulence, the wavelength mismatch between beacon and signal beams fundamentally limits adaptive optics (AO), as both refractive-index accumulation and Fresnel diffraction jointly produce phase discrepancies that cannot be captured by any global scaling law. In high-power regimes, absorption-induced thermal blooming further breaks linear scaling. To address this ill-posed inverse problem, we propose Physics-Aware Cross-Wavelength Network (PAX-Net), a physics-informed network that maps dual-plane beacon intensities at one wavelength to Zernike coefficients at another, embedding differentiable Fresnel transfer-function propagation and phase smoothness constraints as physical priors during training. Simulations demonstrate accurate estimation in both turbulence‐dominated and thermal‐blooming conditions, where conventional scaling fails, confirming that physics‐constrained deep learning offers a viable high‐precision cross‐wavelength wavefront estimation for adaptive optics.