Adaptive Optics Wavefront Lab

An observation-first WebGL console for released adaptive-optics telemetry and optional Noll-index closed-loop experiments.

Adaptive Optics Wavefront Lab boots into recorded adaptive-optics telemetry from ESO CIAO, distributed through the Adaptive Optics Telemetry (AOT) proof-of-concept data release. It shows actual Shack-Hartmann wavefront-sensor streams and high-order deformable-mirror commands. Only after the user enables SIMULATION MODE does the console present generated pupil phase, PID control and point-spread-function products.

Released Telemetry Product

The retained FITS source is:

CIAO1_2019-12-06_DATA_EXPO-015808.fits
https://zenodo.org/records/8192742
doi:10.5281/zenodo.8192742

Zenodo records the file as part of the AOT standard demonstration data, derived from ESO programme 60.A-9278(B) and licensed under Creative Commons Attribution 4.0 International. The local retained source matches the published MD5:

069b37b24997bf55c5312a7bad469502

Its AOT/FITS metadata identifies:

Item	Released value
Instrument	CIAO
AO mode	SCAO
Observation start	2019-12-07T01:58:09.100420 UTC
Guide source	Natural Guide Star
Sensor	Shack-Hartmann
Corrector	High Order Deformable Mirror
Loop state and rate	Closed, 499.962 Hz
Released loop frames	15,000
Header STREHL-R	0.71

The browser asset retains every fiftieth released frame, yielding 300 display samples while keeping the original FITS file locally for provenance and reproducible regeneration. It contains the released AOT extensions:

Extension	Display role	Dimensions in source
GRADIENTS	WFS slope X and Y telemetry maps	15000 x 2 x 68
INTENSITIES	WFS subaperture-flux telemetry map	15000 x 68
HODM POSITIONS	Recorded high-order mirror commands	15000 x 60

These streams are measured telemetry, not reconstructed pupil optical path error. The observation view therefore does not infer an incoming wavefront phase map or a PSF from them.

Instrument Capabilities

• Observation-first WebGL heatmaps for recorded WFS gradient X/Y, WFS subaperture intensity and HODM command streams.
• Recorded telemetry playback with released header Strehl, per-frame slope RMS, command RMS, mean flux and loop-rate readouts.
• Worker-side loading and texture preparation so the UI thread uploads compact display maps.
• Optional simulation mode, disabled by default, for physically declared Noll-index/PID experiments.
• GLSL pupil rendering and worker-side FFT PSF generation in model mode.
• Adjustable proportional and integral gains for servo-lag and baseline-offset experiments.
• Zero-build HTML/CSS/JavaScript deployment.

Architecture

Adaptive Optics Wavefront Lab/
  index.html
  assets/
    css/style.css
    js/app.js                       WebGL presentation and controls
    js/physicsWorker.js             AOT telemetry loading and optional model
  data/observations/
    CIAO1_2019-12-06_DATA_EXPO-015808.fits
    ciao1_aot_telemetry.json        Browser-reduced released streams
    ciao1_aot_metadata.json
  docs/
    equations.md
    validation.md
  tools/
    fetch_aot_observations.py
    validate_observations.py
    validate_model.py
    strehl_table.py

The worker has two explicit paths. In observation mode, it loads and scales only released AOT streams for rendering. In simulation mode, it uses the Noll/PID/FFT model described below. There is no synthetic fallback if the observation asset cannot be loaded.

Optional Noll/PID Model

The generated optical path error uses unit-RMS Zernike polynomials over the circular pupil:

W(rho, theta) = sum_j a_j Z_j(rho, theta)

The controlled Noll modes are:

Noll index	(n, m)	Aberration
J4	(2, 0)	Defocus
J5	(2, -2)	Astigmatism -2
J6	(2, 2)	Astigmatism +2
J7	(3, -1)	Coma -1
J8	(3, 1)	Coma +1
J11	(4, 0)	Primary spherical

For residual modal error e_j, the optional discrete mirror controller is:

I_j[k] = I_j[k-1] + e_j[k] dt
u_j[k+1] = u_j[k] + Kp e_j[k] + Ki I_j[k] + Kd (e_j[k] - e_j[k-1]) / dt

Lower integral gain allows slowly varying residual baseline error to persist; additional latency exposes servo lag. These quantities are model experiments and are not claimed to be the CIAO system controller reconstruction.

The model-mode Strehl estimate is the Marechal approximation for residual RMS in waves:

S approximately exp[-(2 pi sigma_res)^2]

The model PSF panel separately evaluates:

PSF = | FFT2 { P(rho,theta) exp[i 2 pi W_res(rho,theta)] } |^2

Running

Serve this directory over HTTP:

python -m http.server 8000

Open http://localhost:8000/. WebGL is required; no build step or JavaScript dependency installation is required.

To recreate the browser data asset from the official FITS release:

python tools/fetch_aot_observations.py

The converter requires Python, numpy and astropy; a download is required only when the retained source FITS file is absent.

Verification

python tools/validate_observations.py
python tools/strehl_table.py
python tools/validate_model.py

validate_observations.py checks the FITS MD5, release DOI, dimensions and finite telemetry content. The model validators test Zernike normalisation, modal orthogonality and the declared Marechal relationship. These are integrity and numerical checks, not a CIAO PSF reconstruction.

References

Gomes, T., Garcia, P., Correia, C. and Morujao, N. (2023) Proof-of-concept AO telemetry data using the AOT standard format. Zenodo. doi: 10.5281/zenodo.8192742.

Gomes, T. et al. (2024) 'Adaptive optics telemetry standard: Design and specification of a novel data exchange format', Astronomy & Astrophysics, 686, A7. doi: 10.1051/0004-6361/202348486.

Noll, R.J. (1976) 'Zernike polynomials and atmospheric turbulence', Journal of the Optical Society of America, 66(3), pp. 207-211.

Roddier, F. (ed.) (1999) Adaptive Optics in Astronomy. Cambridge: Cambridge University Press.

Licence

Application source is released under the MIT Licence; see LICENSE. The bundled AOT observational product is attributed to its authors and distributed under CC BY 4.0.