Physics-informed machine-learning for modeling aero-optics

We demonstrate the use of physics-informed machine learning algorithms for the adaptive, real-time characterization of aero-optical systems. From deep learning algorithms to nonlinear control methods, the optical sciences are an ideal platform for integrating data-driven control and machine learning for robust characterization and system identification. For the specific case of aero-optics, the ability to extract dominant coherent structures, transients and turbulent behaviors is critical for a diverse number of applications, including the complex and dynamic aero-optic effects on airborne-based laser platforms. Specifically, aero-optical beam control relies on the development of low-latency predictors that can quickly predict aberrated wavefronts to feed into an adaptive optic control loop. We propose develop a number of data-driven methods, including the dynamic mode decomposition (DMD), for real-time forecasting and control.