Machine Learning for Aero-Optical Wavefront Characterization and Forecasting - dissertation

Aug 3, 2026

Laser-based free-space communication faces major obstacles from atmospheric turbulence and fluid-flow-induced refractive-index fluctuations that generate rapidly varying aberrated wavefronts. Aero-optics addresses these effects, enabling adaptive-optics correction to improve coherent transmission, including applications from energy links to secure quantum communications. This dissertation develops low-latency, data-driven predictive methods using machine learning. It presents three techniques—optimized Dynamic Mode Decomposition, shallow-decoder sensor fusion, and recurrent-network wavefront forecasting—to enable robust aero-optical wavefront sensing and phase-distortion forecasting.

LanguageEnglish

CategoryResearch & Report

FormatPDF

Pages100

Reading Time4 h 12 min

AccessFree to download

Abstract

Chapter 1: Aero-Optics

Aero-Optical Effects and Turbulent Boundary Layers

Adaptive Optics

Chapter 2: Sensors

Shack-Hartmann Wavefront Sensor

Digital Holography Wavefront Sensor

Sensor Fusion

Chapter 3: Machine Learning

Dynamic Mode Decomposition

Optimized Dynamic Mode Decomposition

Feed-forward Neural Networks

Recurrent Neural Networks

Chapter 4: Dynamic mode decomposition for aero-optic wavefront characterization

Summary and Introduction

Experimental Data and Data Acquisition

Results and Analysis

Conclusion

Chapter 5: Sensor Fusion for Aero-Optics

Summary and Introduction

Sensor Fusion: Resolution fusion

Sensor Fusion: Multi-modal sensors

Conclusions

Chapter 6: Shallow Recurrent Network Wavefront Forecasting

Summary and Introduction

Methods and Data

Results and Conclusion

Chapter 7: Conclusion

Bibliography

Machine Learning for Aero-Optical Wavefront Characterization and Forecasting Shervin Sahba A dissertation submitted in partial fulfillment of the requirements for the degree of Doctor of Philosophy University of Washington 2023 Reading Committee: J. Nathan Kutz, Chair Arka Majumdar Shih-Chieh Hsu Program Authorized to Offer Degree: Physics ©Copyright 2023 Shervin Sahba University of Washington Abstract Machine Learning for Aero-Optical Wavefront Characterization and Forecasting Shervin Sahba Chair of the Supervisory Committee: J. Nathan Kutz Department of Applied Mathematics The laser is a masterwork of the previous century of physics, but to harness its power coherently in the turbulent wilds of the sky remains a challenge. Free-space lasing hosts myriad applications, from direct energy transmissions for defense to secure communication channels that robustly support quantum entanglement. Aero-optics is the multi-disciplinary field underpinning such optical transmissions through atmosphere and fluid flows. Turbulent variations in the index of refraction, like those forming around boundary layers of airborne optical platforms, manifest aberrated wavefronts. Forecasting these rapid phase distortions allows us to rectify laser transmissions via adaptive optic engineering. Doing so hinges on the development of low-latency predictive techniques, for which we turn to advancements in data-driven algorithms and deep learning. This thesis thus introduces key concepts of aero-optics, wavefront sensing, and machine learning for data-driven physics. We then demonstrate three machine learning methodologies-optimized Dynamic Mode Decomposition, sensor fusion through shallow decoder networks, and forecasting via recurrent neural networks with shallow decoder outputs — for robust aero-optical wavefront sensing. TABLE OF CONTENTS Page List of Figures ....................................... iii List of Tables ........................................ v [Glossary ........................................... vi](Glossary ........................................... vi) [Chapter 1: Aero-Optics ................................](Chapter 1: Aero-Optics ................................). 1 1.1 Aero-Optical Effects and Turbulent Boundary Layers ............. 3 1.2 Adaptive Optics .................................. 4 Chapter 2: Sensors ................................... 7 2.1 Shack-Hartmann Wavefront Sensor ....................... 9 2.2 Digital Holography Wavefront Sensor ...................... 11 2.3 Sensor Fusion ................................... 12 Chapter 3: Machine Learning ............................. 14 3.1 Dynamic Mode Decomposition .......................... 15 3.2 Optimized Dynamic Mode Decomposition ................... 17 3.3 Feed-forward Neural Networks .......................... 18 3.4 Recurrent Neural Networks ............................ 20 Chapter 4: Dynamic mode decomposition for aero-optic wavefront characterization 23 4.1 Summary ..................................... 23 4.2 Introduction .................................... 24 4.3 AAOL-T Experimental Data ........................... 27 4.4 Sensors and Data Acquisition .......................... 28 4.5 Optimized Dynamic Mode Decomposition ................... 29 4.6 Results and Analysis ............................... 32 4.7 Conclusion ..................................... 34 Chapter 5: Sensor Fusion for Aero-Optics ....................... 43 5.1 Summary ..................................... 43 5.2 Introduction .................................... 43 5.3 Aero-Optics .................................... 46 5.4 Sensor Fusion: Resolution fusion ......................... 51 5.5 Sensor Fusion: Multi-modal sensors ....................... 54 5.6 Conclusions .................................... 55 Chapter 6: Shallow Recurrent Network Wavefront Forecasting ........... 62 6.1 Summar

FAQs

Which machine learning methods are introduced for robust wavefront sensing and forecasting?

The dissertation demonstrates three methodologies: optimized Dynamic Mode Decomposition (including feed-forward optimization), sensor fusion using shallow decoder networks, and forecasting with recurrent neural networks with shallow-decoder outputs.

Aug 3, 2026

Upload by Olivia Brown