Any gains in point source detectability or spectroscopic accuracy from data analysis or post-processing ought to be considered in an integrated fashion as part of the overall starlight suppression system.
Multi-wavelength imaging, imaging spectroscopy
Multi-wavelength imaging, imaging spectroscopy
Dual-channel polarimetry
Coherence techniques
Roll or Angular Differencing
Reference Star Subtraction (incl. large Ref. Star Libraries)
Observatory thermal & optical behavior on a variety of timescales
LOWFS and high order WFSC actions on top of those trends
Science instrument properties and detector details
Operational scenarios, target lists, schedules, etc.
From Coronagraph Science Yield Plan presentation Aug 28:
From Coronagraph Science Yield Plan presentation Aug 28:
“Assume systematic noise level, per FWHM of PSF, 1 sigma, to be equal to the photon-counting-limited shot noise of the raw, local speckle intensity”
However, achieving that sensitivity has taken years to decades after launch.
However, achieving that sensitivity has taken years to decades after launch.
For AFTA, we want to be ready with pretty capable methods right from launch.
Sometimes achieving that limit has required developments of special observing modes (e.g. HST spatial scanning spectroscopy, ADI with pupil tracking for AO ).
If this will be the case for AFTA, we need to know that early enough to make sure the necessary observation modes are feasible.
Could assume no subtraction gain at all - conservative, but unrealistic. Risk of not looking like a compelling mission if contrasts are all > 10-8?
Could use a fixed simple estimate - e.g. ”10x” Order of magnitude guess? But not consistent with the level of detail being applied elsewhere.
Could use fundamental photon noise limit - more physically motivated. Similar relative ranking of architectures to using raw contrast curve, scaled by throughput and quadratic vs linear
Could assume some fixed contrast floor we can’t calibrate past - ad hoc without more detailed analyses
Should differences in ability to post process be a discriminator between different coronagraph architectures?
Should differences in ability to post process be a discriminator between different coronagraph architectures?
Perhaps some architectures lead to more “pathological” PSFs than others that are harder to model
Stronger wavelength dependence?
Worse sensitivity to low order aberrations or high order effects like beam walk?
Sensitivity to low order aberrations already specifically included in discriminators
Polarization effects more complicated for VVC
Post-processing for VNCs is not well explored yet
As a practical matter it's hard enough to model these coronagraphs once each during the timeline we have. We're not going to be able to model detailed observing sequences, drifts, reference libraries, systematics.
Using photon noise limit as metric for comparing different architectures is not entirely unreasonable.
Using photon noise limit as metric for comparing different architectures is not entirely unreasonable.
But that’s like a non-constructive existence proof. Doesn’t say anything about how to actually achieve that in practice.
Therefore, over the next few years, conduct an R&D program into PSF subtraction techniques specifically as applied to AFTA
At a level of detail consistent with pre-phase-A, a first pass analysis. Subsequently iterate to higher rigor in later phases of mission development.
Much higher contrast regime. Amplitude as well as phase errors matter.
Much higher contrast regime. Amplitude as well as phase errors matter.
Different thermal stability regime. We don't have any experience operating astronomical telescopes in GEO.
e.g. unexpected "orbital noon" behavior from the initial thermal models?
Speckle correlation timescales and patterns not well known.
Will closing the optimization loop multiple times converge onto different local minima for the dark zone solutions?
How well or not will speckle fields correlate for different stars, or for different spectral bandpasses for one star?
AFTA-related development efforts FY14-16 under discussion with ExEP
AFTA-related development efforts FY14-16 under discussion with ExEP
PISCES IFS prototype under development for HCIT (McElwain et al.) goal to demonstrate very low crosstalk, high dynamic range imaging spectroscopy to accommodate 1e-5 bright halo around 1e-9 dark hole
PISCES IFS prototype under development for HCIT (McElwain et al.) goal to demonstrate very low crosstalk, high dynamic range imaging spectroscopy to accommodate 1e-5 bright halo around 1e-9 dark hole
New proposed activity: Starting in FY14, develop high fidelity optical models and simulator to improve predictions of instrumental systematics for flight AFTA IFS, & provide input simulated data to postprocessing development.
Goal to develop a candidate flight IFS instrument design based on PISCES design
Dual channel (Wollaston, Rochon prisms or micropolarizer grid) vs single channel (wire grid or other linear polarizer)
Dual channel (Wollaston, Rochon prisms or micropolarizer grid) vs single channel (wire grid or other linear polarizer)
Pending input from SDT on prioritization and science drivers for AFTA
Proposed activity: As instrument design and understanding of pol. systematics improve, conduct analyses of pol. contrast gains & science yields; iterate. - VVC needs special consideration
Use estimates of electric field to subtract modeled speckle fields
Use estimates of electric field to subtract modeled speckle fields
Needs further refinement for AFTA.
Combine with iterative WFSC control laws? (Groff, Kasdin, et al.)
Combine with LOWFS telemetry for low order aberrations?
Developed by Ygouf et al. for SPHERE AO project
Developed by Ygouf et al. for SPHERE AO project
Simultaneous retrieval of wavefront aberration, planet locations, and spectra
Overall Goal: Ensure that by FY17, our methods for data analyses are sufficiently mature to aid in demonstrating that the combination of instrument and analyses methods can meet the 1e-9 requirement before project PDR
Overall Goal: Ensure that by FY17, our methods for data analyses are sufficiently mature to aid in demonstrating that the combination of instrument and analyses methods can meet the 1e-9 requirement before project PDR
End FY14: estimation of post processing gain based on existing HCIT data
End FY15: preliminary estimation of post processing gain based on updated simulations and HCIT test data of AFTA coronagraph architecture
End FY16: refined estimation of post processing gain based on simulations and HCIT data
Data analyses methods (postprocessing) need to be considered as an essential part of the overall integrated science system
Data analyses methods (postprocessing) need to be considered as an essential part of the overall integrated science system
This work is just beginning and we do not yet have a mature understanding of the best methods for the 1e-8 to 1e-10 contrast regime.
Photon noise limit is a useful approximation for now, but how to get there?
We need to improve on this over the next few years in parallel with hardware R&D
Assessment of simultaneous differential methods (spectral, polariz.) can start right away
Assessing reference library methods depends on first estimated of timescales and correlations.
Need to start now so we can iterate and feed back lessons to design of the science instrument, spacecraft, and operations scenarios
