Large sparse aperture telescope wavefront sensing and control via pretrained neural network with attention module.

Area of Science:

• Optics and Photonics
• Artificial Intelligence
• Optical Engineering

Background:

• Accurate piston error detection is crucial for co-phasing sparse aperture optical systems.
• Current methods may face limitations in accuracy and range for complex optical setups.

Purpose of the Study:

• To propose a global piston error modulation method for sparse aperture mirrors using convolutional neural networks (CNNs).
• To enhance the accuracy, efficiency, and sensing range of piston error detection.
• To enable precise global fine phase correction in optical systems.

Main Methods:

• Development of a CNN-based global piston error modulation method.
• Integration of a convolutional block attention module (CBAM) with a data generalization mechanism.
• Utilizing actual co-phasing sensor images for training and validation with reduced labeled data.

Main Results:

• The proposed method achieves high prediction accuracy for piston error distribution.
• Demonstrated enhancement in piston error detection efficiency and sensing range.
• Facilitated global fine phase correction to less than λ/80 under closed-loop conditions.

Conclusions:

• The CNN-based method effectively detects piston errors in sparse aperture systems.
• The technique shows significant potential for simplifying wavefront sensing and modulation in large segmented telescopes.
• This approach offers a robust solution for co-phasing challenges in advanced optical systems.