Abstract:We have taken advantage of the implementation of an adaptive optics system on the Themis solar telescope to implement innovative strategies based on an inverse problem formulation for the control loop. Such an approach encompassing the whole system implies the estimation of the pixel variances of the Shack-Hartmann wavefront sensor, a novel real-time method to extract the wavefront slopes as well as their associated noise covariance, and the computation of pseudo-open loop data. The optimal commands are computed by iteratively solving a regularized inverse problem with spatio-temporal constraints including Kolmogorov statistics. The latency of the dedicated real-time control software with conventional CPU is shorter than 300 $\mu$s from the acquisition of the raw 400 x 400 pixel wavefront sensor image to the sending of the commands.

What problem does this paper attempt to address?

The problem that this paper attempts to solve is: How to implement innovative real - time control strategies in the solar Adaptive Optics (AO) system to improve the accuracy and speed of wave - front correction. Specifically, by formulating the control - loop problem as an inverse problem, the author proposes new methods to estimate the pixel variance of the Shack - Hartmann wave - front sensor, extract the wave - front slope and its noise covariance matrix in real - time, and calculate the pseudo - open - loop data. These methods aim to optimize the commands applied to the Deformable Mirror (DM), thereby correcting the wave - front distortion caused by atmospheric turbulence more effectively. ### Key Technical Points 1. **Formulation of Inverse Problem**: The author uses the inverse - problem approach to solve the control problem of the adaptive optics system. This method requires estimating the pixel variance of the wave - front sensor and has developed new real - time methods to extract the wave - front slope and its noise covariance matrix. 2. **Pseudo - Open - Loop Data**: The paper proposes using pseudo - open - loop data for control. This requires a good system interaction - matrix model, which is calibrated by measuring the effects of sending random commands. 3. **Iterative Solution of Regularized Inverse Problem**: The optimal commands are calculated by iteratively solving the regularized inverse problem with spatio - temporal constraints, including Kolmogorov statistical prior information. 4. **Low - Latency Real - Time Control Software**: The real - time control software runs on a conventional CPU, and the delay time from obtaining the original 400×400 - pixel wave - front sensor image to sending commands is less than 300 microseconds. ### Formula Summary - Linearized model of wave - front sensor data: \[ d_t = S\cdot(w_t + M\cdot a_t)+z_t \] where $d_t$ is the wave - front sensor data, $S$ is the linear response matrix of the sensor, $w_t$ is the wave - front, $M$ is the linear response matrix of the deformable mirror (i.e., the influence matrix), $a_t$ is the actuator command, and $z_t$ is the noise. - Calculation of optimal commands: \[ a_{t + \delta t}=-M^\dagger\cdot w_{t + \delta t|t} \] where $M^\dagger=(M^T\cdot M)^{-1}\cdot M^T$ is the pseudo - inverse of the linear response matrix of the deformable mirror, and $w_{t + \delta t|t}$ is the wave - front expectation at time $t + \delta t$ given the information at time $t$. - Pseudo - open - loop data: \[ y_t = d_t - G\cdot a_t=S\cdot w_t+z_t \] where $G = S\cdot M$ is the interaction matrix. - Update rule: \[ x|_t=x|_{t - 1}+A_{t|t - 1}\cdot(w_{t|t}-w_{t|t - 1}) \] Through these methods, the paper shows how to achieve efficient real - time control in the solar adaptive optics system, thereby significantly improving the observation quality.

Closing the loop as an inverse problem: the real-time control of THEMIS adaptive optics

The wavefront sensing making-of for THEMIS solar telescope

Efficient implementation of pseudo open loop control for adaptive optics on Extremely Large Telescopes

The future looks dark: improving high contrast imaging with hyper-parameter optimization for data-driven predictive wavefront control

Open loop calibration and closed loop non-perturbative estimation of the lateral errors of an adaptive optics system: examples with GRAVITY+ and CHARA experimental data

Data-driven Subspace Predictive Control of Adaptive Optics for High-Contrast Imaging

Modeling astronomical adaptive optics performance with temporally filtered Wiener reconstruction of slope data

Integrated modeling of wavefront sensing and control for space telescopes utilizing active and adaptive optics

Identification and Adaptive Control of a High-Contrast Focal Plane Wavefront Correction System

Real-time Adaptive Optics with pyramid wavefront sensors: A theoretical analysis of the pyramid sensor model

Closed-Loop Until Further Notice: Comparing Predictive Control Methods in Closed-Loop

Advances in control of a Pyramid Single Conjugate Adaptive Optics system

Revisiting the Borde-Traub focal plane wavefront estimation technique for exoplanet direct imaging

Modelling astronomical adaptive optics performance with temporally-filtered Wiener reconstruction of slope data

Apollon Real-Time Adaptive Optics (ARTAO) -- Astronomy-Inspired Wavefront Stabilization in Ultraintense Lasers

LUVOIR-ECLIPS closed-loop adaptive optics performance and contrast predictions

Self-optimizing adaptive optics control with Reinforcement Learning for high-contrast imaging

Forecasting Wavefront Corrections in an Adaptive Optics System

Estimation of the lateral mis-registrations of the GRAVITY + adaptive optics system

Wavefront Sensing and Control with the Many Headed Hydra Modal Basis