Finite element-wavelet hybrid algorithm for atmospheric tomography.

Area of Science:

• Astronomy and Astrophysics
• Computational Science
• Optical Engineering

Background:

• Atmospheric tomography is essential for adaptive optics (AO) systems in next-generation telescopes like the European Extremely Large Telescope (E-ELT).
• Traditional matrix-vector multiply (MVM) methods face computational challenges with the high dimensionality of these problems.
• Previous work introduced a wavelet-based conjugate gradient algorithm for atmospheric tomography.

Purpose of the Study:

• To investigate and improve the computational efficiency of wavelet-based algorithms for atmospheric tomography.
• To develop a new finite element-wavelet hybrid algorithm that is globally O(n), parallelizable, and memory-compact.
• To evaluate the performance and quality of the new algorithm compared to existing methods.

Main Methods:

• Introduced three novel techniques: dual domain discretization, a scale-dependent preconditioner, and a ground layer multiscale method.
• Developed a finite element-wavelet hybrid algorithm integrating these techniques.
• Estimated computational costs and compared theoretical performance with MVM methods.
• Evaluated the algorithm's quality using a multiobject adaptive optics (MOAO) simulation for the E-ELT within the OCTOPUS simulation system.

Main Results:

• The developed finite element-wavelet hybrid algorithm is globally O(n), parallelizable, and memory-compact.
• Computational cost estimates suggest significant improvements over traditional MVM methods.
• Simulations demonstrate the method's effectiveness and quality for MOAO systems.

Conclusions:

• The new wavelet-based hybrid algorithm offers a computationally efficient and scalable solution for atmospheric tomography.
• This method is well-suited for the demands of next-generation AO systems, particularly for extremely large telescopes.
• The developed techniques provide a robust alternative for real-time atmospheric turbulence reconstruction.