Neural-network-based wavefront solution algorithm for a wide field survey telescope.

Area of Science:

• Astronomy and Astrophysics
• Optical Engineering
• Computational Science

Background:

• The Wide Field Survey Telescope (WFST) is a 2.5m optical survey telescope utilizing a primary-focus design for a 3-degree field of view.
• Effective wavefront sensing and active optics are crucial for the WFST's performance, relying on curvature sensors.
• Existing wavefront solution algorithms (FFT, orthogonal series, Green's function, sensitivity matrix) have practical limitations.

Purpose of the Study:

• To propose and evaluate a novel convolutional neural network (CNN) based solution for curvature wavefront sensing.
• To address the limitations of current wavefront sensing algorithms for the WFST.
• To demonstrate the efficacy and superiority of the proposed CNN method.

Main Methods:

• Development of a U-Net structured convolutional neural network model for curvature wavefront sensing.
• Training and numerical simulation of the CNN model using WFST-relevant data.
• Comparative analysis of the CNN method against the established sensitivity matrix method.

Main Results:

• The trained U-Net CNN model achieves high accuracy in curvature wavefront solutions.
• The proposed method effectively performs wavefront sensing for the WFST.
• Numerical simulations indicate the CNN approach outperforms the sensitivity matrix method.

Conclusions:

• The U-Net CNN model offers a superior and accurate solution for curvature wavefront sensing in the WFST.
• This AI-driven approach enhances the capabilities of active optics systems in large survey telescopes.
• Future work will focus on addressing the identified drawbacks and further optimizing the CNN model.