Master's Thesis Defense in Physics: Michael Jones 4/8

03/20/2025
By Michael Jones

The Kennedy College of Sciences, Department of Physics and Applied Physics, invites you to attend a Master’s thesis defense by Michael Jones on "Adaptive Optics Wavefront Stabilization Using a Convolutional Neural Network."

Defense Date: Tuesday, April 8
Time: 10 a.m. - 12 p.m.
Location: Wannalancit, Room 405

Contact Michael Jones at Michael_Jones6@student.uml.edu for a Zoom link if needed.

Advisor: Christopher Mendillo, Ph.D.
Committee Members:
- Timothy Cook, Ph.D.
- Caroline Desgranges, Ph.D.

Abstract:
The Planetary Imaging Coronagraph Testbed Using a Recoverable Experiment for Debris Disks (PICTURE-D) balloon mission aims to image the debris disk in several nearby exoplanetary systems. To achieve the required starlight suppression of 10^-7, PICTURE-D makes use of a vector vortex coronagraph and three deformable mirrors. The reflective Lyot stop Low-Order Wavefront Sensor (LLOWFS), which includes a camera and one of the deformable mirrors, is responsible for high-speed stabilization of the incoming wavefront. In this thesis, a convolutional neural network (CNN) is employed to determine the Zernike coefficients present in the wavefront via a single, out-of-focus image taken on the LLOWFS. It is shown that the CNN is able to meet the mission requirements for speed and accuracy when run on the low-power flight computer. Additionally, it is shown that the CNN has much lower cross-coupling between Zernike modes and supports a much higher range of aberrations than the previous linear estimator used on the LLOWFS.