AO4ELT 2

Quasi-real-time end-to-end simulations of ELT-scale adaptive optics systems on GPUs

Damien GRATADOUR — 2011-08-19T10:05:40Z

Submitted by D. Gratadour

Authors

Damien Gratadour

Affiliations

LESIA

Abstract

Our team has started the development of a code dedicated to GPUs for the simulation of AO systems at the E-ELT scale. It uses the CUDA toolkit and an original binding to Yorick (an open source interpreted language) to provide the user with a comprehensive interface.

In this paper we present the first performance analysis of our simulation code, showing its ability to provide Shack-Hartmann (SH) images and measurements at the kHz scale for VLT-sized AO system and in quasi-real-time (up to 70 Hz) for ELT-sized systems on a single top-end GPU. The simulation code includes multiple layers atmospheric turbulence generation, ray tracing through these layers, image formation at the focal plane of every sub-apertures of a SH sensor using either natural or laser guide stars and centroiding on these images using various algorithms. Turbulence is generated on-the-fly giving the ability to simulate hours of observations without the need of loading extremely large phase screens in the global memory. Because of its performance this code additionally provides the unique ability to test real-time controllers for future AO systems under nominal conditions.

Integrating AO in a performance budget: towards a global system engineering vision

Philippe LAPORTE — 2011-05-19T21:23:49Z

Submitted by P. Laporte

Authors

Philippe Laporte (a) and Hermine Schnetler (b)

Affiliations

(a) GEPI, (b) UK ATC

Abstract

EAGLE (Extremely large Adaptive telescope for GaLaxy Evolution) is one of the eight E-ELT instruments concepts that was developed as part of the Phase A E-ELT instrument studies. EAGLE is a near-infrared wide field multi object spectrograph. It includes its own multi-object adaptive optics system (MOAO) and its subsystems are cooled down so as to ensure that the instrument can both achieve the desired spatial resolution in the K-band and to ensure that the instrument is background limited, as required in the primary science case. In this paper we describe the method we put in place to partition and allocate the important characteristics to the various subsystems and how the adaptive optics improves the overall performance. We also discuss the process which verify that the concept design will deliver the required characteristics. Due to the integrated nature of the instrument, a large number of AO parameters have to be controlled. The performance matrix also has to deal with the added complexity of active optical elements such as the science channel deformable mirrors (DMs). This paper also defines a method of how to convert the ensquared energy (EE) and signal-to-noise ratio (SNR) required by the primary science case into the “as designed” wavefront error and the overall residue wavefront error. To ensure successful integration and verification of the next generation instruments for ELT it is of the utmost importance to have method to control and managed the instrument's critical performance characteristics using adaptive optics.

Laser-Guide Star Point-Spread Function Reconstruction for ELTs

Carlos CORREIA — 2011-05-17T15:01:41Z

Submitted by C. Correia

Authors

C. Correia (a) , J.-P. Véran (a), B. Ellerbroek (b), L. Gilles (b) and L. Wang (b)

Affiliations

ONERA

Abstract

To exploit the maximum potential of Extremely Large Telescopes (ELT), adaptive-optics (AO)-corrected images can be further enhanced by using image restoration techniques. Such techniques rely on accurate knowledge of the point-spread function (PSF) anywhere in the field.

To increase sky-coverage ELTs use laser beacons to probe the three-dimensional atmosphere and multi-conjugate AO to increase correction above the isoplanatic patch. These features translate into three sources of anisoplanatism: 1) focal anisoplanatism known as cone effect 2) angular anisoplanatism due to the difference of the wave-fronts in the LGS and science directions and 3) tip/tilt angular anisoplanatism, on account of the LGS being blind to TT, the latter being estimated from natural GS measurements in different locations in the field.

In our approach, the long-exposure science optical transfer function (OTF) (Fourier transform of the PSF) is estimated as a product of 3 terms: (i) the OTF of a point-source LGS, estimated from system telemetry using Véran's method, (ii) a model anisoplanatism filter, computed from a high-fidelity numerical simulation to account for the difference between the OTFs for a point-source at the location of the LGS and the science target, and (iii) a tip/tilt/tilt anisoplanatism filter obtained from system and model telemetry, and expressed as a system-to-model OTF ratio.

We present the first stage of the reconstruction, (ii) and (iii) being presented by Gilles in this same conference. Based on [Véran et al 97], we show how to reconstruct the LGS-PSF by de-noising telemetry-accumulated measurements and removing AO-loop specific terms, namely the measurement noise and aliasing components; we present the modifications needed on account of the system size and the optimisations required to accurately reconstruct the PSF. Furthermore, we compare our estimates to those of a high-fidelity Monte-Carlo simulator (MAOS) that can accurately model the PSF in those locations.

Simulations of a SCAO System for an E-ELT telescope using the pyramid wavefront sensor: recent results

Aurea GARCIA RISSMANN — 2009-02-28T23:00:00Z

Submitted by Aurea GARCIA RISSMANN

Authors

Aurea Garcia-Rissmann, Miska Le Louarn (1)

Affiliations

(1) European Southern Observatory

Abstract

Updated results obtained from simulations of a single-conjugated adaptive optics (SCAO) system for a 42m telescope are presented in this work. The modulated pyramid wavefront sensor (PWFS), which has been preferred in some projects for its versatility and sensitivity, is used in these simulations. The main objective of this work is to evaluate the performance of such SCAO system under different parameters (loop gain, modulation, truncated SVD mode), sensing wavelengths, atmospheric coherence scales and NGS magnitudes. Always measuring the Strehl ratio in the K-band, we have verified that the overall performance tends to be poorer as the sensing wavelength becomes shorter. The loop gain optimal range is dependent on the SVD truncation threshold used to build the command matrix, and a non-modulated PWFS produces in general poorer results when compared to modulated cases, being this especially true for the R-sensing band. The default atmospheric model adopted was a von Karman with r0=0.13m (at 500nm) and outer scale of 25m, but poorer and better seeing conditions have also been tested. We also show how the Strehl is affected by the incidence of different photon fluxes at the PWFS detector.

PSF reconstruction for NICI, the high-contrast coronagraphic imager of GEMINI observatory

Markus HARTUNG — 2009-02-28T23:00:00Z

Submitted by Hartung MARKUS

Authors

M. Hartung (1), D. Gratadour (2), M. Chun (3), T. Hayward (1)

Affiliations

(1) Gemini Observatory, (2) Observatoire Paris-Meudon, (3) Institute for Astronomy, Hawaii, USA

Abstract

It is more than a decade ago that PSF reconstruction from wavefront sensor (WFS) data has been described by Veran et al. (JOSA A, 1997) and successfully demonstrated at CFHT/PUEO. Nevertheless, even though adaptive optics (AO) has evolved into a mature technology, no breakthrough has been reached yet in terms of a broad use of WFS estimated PSFs in astronomical image data reduction. Reasons for this are certainly the lack of easy access to the needed AO data as well as the automatic supply of reconstructed PSFs along with the science exposures to eventually give the astronomical community a chance to exploit its potential. The key information are the covariance matrices of the WFS signal and the DM voltages in sync with the science exposures. Without active interaction from the user side with the observatory, we intend to automatically provide the user with these AO data together with the retrieved PSF for every taken science exposure.

Typical applications for a WFS retrieved PSF are photometry and image sharpening by deconvolution. For the high-contrast imager NICI, the reconstructed PSF will be of particular use providing a fixed reference point to derive planet detection contrast curves completely in sync with the science exposure. We will compare how well the WFS estimated PSF matches the science image PSF at various AO guide stars, and will study the impact of system modeling errors.

Raven performance modeling

Meguru ITO — 2009-02-28T23:00:00Z

Submitted by Meguru ITO

Authors

Meguru Ito, Kate Jackson, David Anderson, Raven team

Affiliations

University of Victoria, NRC Herzberg Institute of Astrophysics

Abstract

Raven is a MOAO demonstrator that will be used on the Subaru telescope and will hopefully serve as a pathfinder for future ELT MOAO instruments. In this paper, we present trade studies and the resultant error budget that were used to define the AO architecture for Raven. In particular, we focus here on recent simulations of Raven performance for real science targets, and performance as a function of wavelength, zenith angle and atmospheric conditions. We show that Raven should achieve 30% ensquared energy within 140 mas under median conditions under realistic observing conditions.

Analytical vs. end-to-end numerical modeling of adaptive optics systems: comparison between PAOLA and the Software Package CAOS.

Marcel CARBILLET — 2009-02-28T23:00:00Z

Submitted by Marcel CARBILLET

Authors

Marcel Carbillet [1] & Laurent Jolissaint [2]

Affiliations

[1] : UMR 6525 H. Fizeau (UNS/CNRS/OCA) [2] : aquilAOptics

Abstract

We compare in this contribution the analytical approach together with the so-called "end-to-end" approach in the framework of astronomical adaptive optics (AO) modeling. The two tools used for this purpose are well-known and already widely used within the astronomical AO community: PAOLA (Jolissaint et al. 2006, Jolissaint 2010) on the one hand, and the Software Package CAOS (Carbillet et al. 2005) on the other hand. In addition to inter-validate the two codes, trade-offs are clearly searched in order to find optimal compromises permitting to face both exploratory simulations and large instrumental projects while combining effectiveness and certainty.

An Iterative Model for Micro-Electro-Mechanical-Systems (MEMS) Deformable Mirrors

Célia BLAIN — 2009-02-28T23:00:00Z

Submitted by Célia BLAIN

Authors

Célia Blain(1), Olivier Guyon(2), Colin Bradley(1), Frantz Martinache(2), Christophe Clergeon(2)

Affiliations

(1) University of Victoria Adaptive Optics Laboratory (2) Subaru Telescope, NAOJ

Abstract

We present a high accuracy Micro-Electro-Mechanical-System (MEMS) deformable mirror (DM) control algorithm which was implemented in the real-time control interface of the Subaru Coronagraphic Extreme Adaptive Optics project (SCExAO) for the on-sky engineering run of July 2011. MEMS DMs are an attractive DM technology for ExAO because they offer unprecedented actuator density and actuator counts. The small size and low cost per actuator also make them attractive for Multi-Object AO (MOAO), where one DM is needed per science field. These applications both require a DM model capable of reproducing a phase map with a precision of few nm rms. The high accuracy DM control algorithm presented below is suitable for open-loop AO systems where DM calibration is essential. For SCExAO, the high accuracy wavefront sensing uses DM modulations to coherently mix light with slow/static speckles and precisely measure their complex amplitude. Therefore, wavefront sensing and correction both depend on precise DM control. The model (described in Blain et al., 2011) which relies (i) on a physical model of the actuators and the membrane and (ii) on the optimization of 8 coefficients, could be adopted as an open-loop control solution for future MOAO or ExAO ELTs instruments. During the initial test phase at the UVic AO Lab, the performance of the model reached an open-loop error equal to 7.3% of the rms of the desired phase (1.6% of the peak-to-valley(PV) of the desired phase) with Kolmogorov type wavefront (test performed over 10 phase screens with a mean PV of 1448.3 nm and a mean rms of 489.5 nm).

Aberrations induced by side-projected laser guide stars in laser tomography adaptive optics systems

Marcos VAN DAM — 2009-02-28T23:00:00Z

Submitted by Marcos VAN DAM

Authors

Marcos A. van Dam, Rodolphe Conan, Antonin H. Bouchez and Brady Espeland

Affiliations

(1) Flat Wavefronts (2) and (4) Australian National University (3) Giant Magellan Telescope Observatory

Abstract

A laser tomography adaptive optics (LTAO) system is currently under design for the Giant Magellan Telescope (GMT). For systems engineering reasons, it is preferable to project the laser guide stars (LGSs) from the side of the telescope. Experience with the Keck II adaptive optics system and analytical modeling have shown that side-launched lasers result in aberrations, called LGS aberrations, with a lot of power at low-spatial frequencies. This is caused by the elongation of the LGS due to the finite thickness of the sodium layer.

In this paper, we model the LGS aberrations for the GMT's LTAO system, which has three launch telescopes projecting six LGSs in a regular hexagon. When the wavefront is reconstructed tomographically, the aberrations largely cancel on-axis. However, an off-axis truth sensor with a dedicated DM on the truth sensing path will see completely different aberrations, introducing an unacceptably large wavefront error. Working with the assumption that the LGS aberrations affect all of the wavefront sensors in the sane way, we propose a method that filters the LGS aberrations directly from the Shack-Hartmann centroids. This method is shown to be very effective at filtering the LGS aberrations while having a negligible effect on the turbulence estimation.

Study of GLAO-corrected PSF evolution for the MUSE Wide Field Mode. Expected performance and requirements for PSF reconstruction

Thierry FUSCO — 2009-02-28T23:00:00Z

Submitted by Thierry FUSCO

Authors

T. Fusco(1), R. Villecroze(2), A. Jarno(2), R. Bacon(2)

Affiliations

(1) ONERA / DOTA (2) Observatoire de Lyon / CRAL

Abstract

The second generation instrument MUSE for the VLT has been designed to profit of the ESO Adaptive Optics Facility (AOF). The two Adaptive Optics (AO) modes (GLAO in Wide Field Mode [WFM] and LTAO in Narrow Field Mode [NFM]) will be used. To achieve its key science goals, MUSE will require information on the full system (Atmosphere, AO, telescope and instrument) image quality and its variation with Field position and wavelength. For example, optimal summation of a large number of deep field exposures in WFM will require a good knowledge of the PSF.

In this paper, we will present an exhaustive analysis of the MUSE Wide Field Mode PSF evolution both spatially and spectrally. For that purpose we have coupled a complete AO simulation tool developed at ONERA with the MUSE instrumental PSF simulation. Relative impact of atmospheric and system parameters (seeing, Cn², LGS and NGS positions etc …) with respect to differential MUSE aberrations per channel (i.e. slicer and IFU) is analysed. The results allow us (in close collaboration with astronomers) to define pertinent parameters (fit parameters using a Moffat function) for a PSF reconstruction process (estimation of this parameters using GLAO telemetry) and to propose an efficient and robust algorithm to be implemented in the MUSE pipeline.

The extension of the spatial and spectral PSF analysis to the NFM case is discussed and preliminary results are given. Some specific requirements for the generalisation of the GLAO PSF reconstruction process to the LTAO case are derived from these early results.