Impact of time-variant turbulence behavior on prediction for adaptive optics systems

Kooten, Maaike A. M. van; Doelman, Niek; Kenworthy, Matthew A.

doi:10.1364/josaa.36.000731

Cited by 25 publications

(16 citation statements)

References 31 publications

(36 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Instead of creating a single filter whose inputs and outputs include all 349 actuator positions, we now create a different filter for each individual actuator. The history vector now contains the actuator positions from either a 3 × 3 or a 5 × 5 box centered on the actuator of interest (see also the approach described in van Kooten et al 2019 17 ). For simplicity, we disregard edge actuators where a complete box is not possible.…”

Section: Exploring the Spatial Filter Parametersmentioning

confidence: 99%

Demonstrating predictive wavefront control with the Keck II near-infrared pyramid wavefront sensor

Jensen-Clem

Bond

Cetre

et al. 2019

Techniques and Instrumentation for Detection of Exoplanets IX

View full text Add to dashboard Cite

The success of ground-based instruments for high contrast exoplanet imaging depends on the degree to which adaptive optics (AO) systems can mitigate atmospheric turbulence. While modern AO systems typically suffer from millisecond time lags between wavefront measurement and control, predictive wavefront control (pWFC) is a means of compensating for those time lags using previous wavefront measurements, thereby improving the raw contrast in the post-coronagraphic science focal plane. A method of predictive control based on Empirical Orthogonal Functions (EOF) has previously been proposed and demonstrated on Subaru/SCExAO. In this paper we present initial tests of this method for application to the near-infrared pyramid wavefront sensor (PYWFS) recently installed in the Keck II AO system. We demonstrate the expected root-mean-square (RMS) wavefront error and contrast benefits of pWFC based on simulations, applying pWFC to on-sky telemetry data saved during commissioning of the PYWFS. We discuss how the performance varies as different temporal and spatial scales are included in the computation of the predictive filter. We further describe the implementation of EOF pWFC within the PYWFS dedicated real-time controller (RTC), and, via daytime testing at the observatory, we demonstrate the performance of pWFC in real time when pre-computed phase screens are applied to the deformable mirror (DM).

show abstract

Section: Exploring the Spatial Filter Parametersmentioning

confidence: 99%

Demonstrating predictive wavefront control with the Keck II near-infrared pyramid wavefront sensor

Jensen-Clem

Bond

Cetre

et al. 2019

Techniques and Instrumentation for Detection of Exoplanets IX

View full text Add to dashboard Cite

show abstract

“…One of the largest challenges of the model-based predictive controllers is the effect of changes in model parameters such as wind speed and direction. 32 Therefore, it is crucial to test the algorithm Fig. 7 The radial profiles of the post-coronagraphic contrast maps.…”

Section: Non-stationary Turbulencementioning

confidence: 99%

“…The second set of simulations shows that even completely random changes in the conditions can be handled. The actual wind speed has more temporal structure, 32 therefore, we expect that better performance can be achieved with more realistic wind models. 4 Low-Order Control Verification with MagAO-X…”

Section: Non-stationary Turbulencementioning

confidence: 99%

“…This is limiting the application of open-loop predictive controllers. 32 Another complication stems from the use of multi-stage AO systems, where several dynamical components are operated at the same time, but each may be operated at different frequencies. The temporal dynamics of each component need to be accurately accounted for to reconstruct the full open-loop disturbance.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Data-driven subspace predictive control of adaptive optics for high-contrast imaging

Haffert

Males

et al. 2021

J. Astron. Telesc. Instrum. Syst.

View full text Add to dashboard Cite

The search for exoplanets is pushing adaptive optics (AO) systems on ground-based telescopes to their limits. One of the major limitations at small angular separations, exactly where exoplanets are predicted to be, is the servo-lag of the AO systems. The servo-lag error can be reduced with predictive control where the control is based on the future state of the atmospheric disturbance. We propose to use a linear data-driven integral predictive controller based on subspace methods that are updated in real time. The new controller only uses the measured wavefront errors and the changes in the deformable mirror commands, which allows for closed-loop operation without requiring pseudo-open loop reconstruction. This enables operation with non-linear wavefront sensors such as the pyramid wavefront sensor. We show that the proposed controller performs nearoptimal control in simulations for both stationary and non-stationary disturbances and that we are able to gain several orders of magnitude in raw contrast. The algorithm has been demonstrated in the lab with MagAO-X, where we gain more than two orders of magnitude in contrast.

show abstract

“…Another issue for these approaches is non-stationary turbulence. 20 All these methods require online updating of the control parameters based on the current turbulence parameters. This updating of the control law is computationally very expensive and often involves inverting large matrices.…”

Section: Introductionmentioning

confidence: 99%

Self-optimizing adaptive optics control with reinforcement learning

Landman

Haffert

Radhakrishnan

et al. 2020

Adaptive Optics Systems VII

View full text Add to dashboard Cite

Current and future high-contrast imaging instruments require extreme Adaptive Optics (XAO) systems to reach contrasts necessary to directly image exoplanets. Telescope vibrations and the temporal error induced by the latency of the control loop limit the performance of these systems. Optimization of the (predictive) control algorithm is crucial in reducing these effects. We describe how model-free Reinforcement Learning can be used to optimize a Recurrent Neural Network controller for closed-loop adaptive optics control. We verify our proposed approach for tip-tilt control in simulations and a lab setup. The results show that this algorithm can effectively learn to suppress a combination of tip-tilt vibrations. Furthermore, we report decreased residuals for power-law input turbulence compared to an optimal gain integrator. Finally, we demonstrate that the controller can learn to identify the parameters of a varying vibration without requiring online updating of the control law. We conclude that Reinforcement Learning is a promising approach towards data-driven predictive control; future research will apply this approach to the control of high-order deformable mirrors.

show abstract

Impact of time-variant turbulence behavior on prediction for adaptive optics systems

Cited by 25 publications

References 31 publications

Demonstrating predictive wavefront control with the Keck II near-infrared pyramid wavefront sensor

Demonstrating predictive wavefront control with the Keck II near-infrared pyramid wavefront sensor

Data-driven subspace predictive control of adaptive optics for high-contrast imaging

Self-optimizing adaptive optics control with reinforcement learning

Contact Info

Product

Resources

About