KBO encounters Options

[SteveM]

To clarify the previous items, they all reflect the state of discussion in 2010-2011, enroute to Pluto. The items cited are:

Alan Stern, [re LORRI]: Apr 15 2010, 07:18 AM

John_S, [re Science at the Centaurs]: Apr 28 2010, 02:36 PM

John_S, [re KBOs]: Feb 22 2011, 05:55 PM

Dates are important in citations, especially when the problems are evolving.

Steve M

[HSchirmer]

To clarify the previous items, they all reflect the state of discussion in 2010-2011, enroute to Pluto.

I've added the dates to clarify.

Can be a bit distracting flipping between two threads on the same topic...

http://www.unmannedspaceflight.com/index.p...st&p=243214

[Alan Stern]

I've added the dates to clarify.

Can be a bit distracting flipping between two threads on the same topic...

http://www.unmannedspaceflight.com/index.p...st&p=243214

Note we will look at using JWST as well, maybe HST, and then eventually the 25-30 m ELTs coming on line. Until we find something we are going to throw everything there is at it.

[scalbers]

Note we will look at using JWST as well, maybe HST, and then eventually the 25-30 m ELTs coming on line. Until we find something we are going to throw everything there is at it.

I wonder if LSST will go deep enough with its rapid search capability, at least for larger potential objects. This may also represent a citizen science opportunity with all the data being made public.

[HSchirmer]

This may also have a citizen science opportunity with all the data being made public.

Interesting point about citizen science, wonder if there's any precedent for a "white hat" programming competition at Standford with a prize for any team that one can write a stable "KBO image identification routine" that will run in the background on NH's backup CPU?

Yes, that sounds crazy at first, but remember
NH is running a PlayStation CPU.
and
There are VMware PlayStation emulators to play PS1 games on PCs.

[WTW]

I wonder if LSST will go deep enough with its rapid search capability, at least for larger potential objects. This may also represent a citizen science opportunity with all the data being made public.

https://www.lsst.org/science/solar-system/oss

The HST observations of UT by Buie and Parker et al went below ~27.5 magnitude, but that was a "pencil beam" search over many orbits. Hubble's WFC3 UVIS camera has an angular resolution of 0.04 arc sec per pixel (0.13 arcmin in the IR).

LSST will go down to about magnitude ~24.5 with each pair of 15-second exposures per filter; certain areas of the sky with special very deep coverage (<10% observing time) will be able to detect objects at r = ~26. Angular resolution will be ~0.2 arc sec under ideal seeing conditions.

That's probably below the limit of being able to detect Ultima Thule (magnitude ~27.5), for example, but could certainly find things that are brighter (i.e., typically larger). They expect to find and track up to 40,000 TNOs over their first 10 years.

If viewing conditions permit, each LSST field will be imaged twice per night, at intervals between 15 and 60 minutes between "visits". As well as the entire visible southern sky, the survey will include areas within 10 degrees of the northern ecliptic with certain filters when it's visible from northern Chile.

But there are issues:
The normal imaging cadence will exclude a region ~1000 square degrees around the galactic center; the high stellar density and high luminance leads to saturation of the sensors and confusion of the automated system. (Even the motion of the brightest asteroids will not be catalogued, due to saturation issues.)
Exposures in this region will be less frequent and with fewer filter bands.

There are also potential issues with the way that the LSST (and the MPC) will track and report on "moving objects": they plan to identify things that move from night to night, but a small distant KBO may appear to not move enough in 24 hours (to shift position relative to background stars) to be detected as "moving" at the resolution of the LSST optics.
Doing frequent searches over multiple nights becomes computationally very expensive.

But the main problem is that it probably won't start science observations until 2023 at the earliest.

Perhaps someone could weigh in on how all that would affect New Horizons target observations.


For comparison:
The other sky surveys currently in operation aren't really well suited for this kind of search:
The Zwicky Transient Facility at Palomar can only go down to ~21 magnitude with good seeing, and with only an angular resolution of 2 arc seconds. (It was designed to detect brighter transients that don't require only the largest of telescopes to do quick follow-up observations.)

TESS is designed to look for transients of objects that are relatively close and bright: an angular resolution of about 0.35 arcmin and brighter than about magnitude 14 in certain filter bands.

NEOWISE can detect objects in the IR at magnitude of ~15 with an angular resolution of ~ 5 arc sec.

Other ground-based sky surveys use multiple smaller telescopes at various locations; relatively low sensitivity, varying seeing conditions and calibration issues aren't well suited for detecting faint KBOs.

The Giant Magellan Telescope won't be operational until after 2025.

It took a special request to the Gaia team to obtain the information necessary to be able to calculate the occultation data used to accurately track and measure UT. Will such data be in future Gaia data releases?

That leaves JWST, and Hubble -- and getting adequate observing time on them.

So this stuff is challenging! Is there something I'm missing?

[stevesliva]

Quick recap from prior discussions

Right, and the most recent, starting here:
http://www.unmannedspaceflight.com/index.p...st&p=241744

The more recent press conference discussed starfield subtraction from LORRI images, and they look a heckuva lot like the KBO zoo images from back then.

My recollection of an answer about future plans from the press conference is that there's a hope to use nothing but LORRI, but it sounded like the team was still assessing feasibility.

[pioneer]

That leaves JWST, and Hubble -- and getting adequate observing time on them.

So this stuff is challenging! Is there something I'm missing?

I'm a little confused why the JWST is being mentioned when it isn't scheduled to be launched until 2021. https://jwst.nasa.gov/about.html

[HSchirmer]

So this stuff is...
challenging! Is there something I'm missing?

Good summary -
What is often missing is color: not just the visible colors, but the colors we can't see-
Ground-based visible-light telescopes looking for KBOs against the galactic center?
Hmm, that's a bit like somebody looking at the NYC skyline during New Years Eve fireworks, and trying to resolve a charcoal brickette placed on the top of the times-square-ball...
Looking in the infrared, a frequency where KBOs are NOT incredibly dark, has some benefits...

Also, IIRC, the prior discussions about LORRI looking for KBOs actually dealt with 3 different arrangements, (and was not always clear about which arrangement was being discussed).
(greatly oversimplifying)
1) LORRI was designed to take photos while slewing sideways in thruster-mode as it passes a target, e.g. Jupiter/Pluto/UT
2) LORRI can also stare straight ahead to take navigation photos in thruster-stabilised-mode.
- Problem is both of those use up thruster propellant reasonably quickly.

3) Question was, whether LORRI could save fuel by taking photos in spin-stabilized mode, i.e. spinning with the big dish pointing toward Earth for good communication, while constantly looking "sideways" into the Kuiper Belt for random nearby KBOs.
The answer was generally, nope. With no idea of where and when to take pictures, long exposures would accumulate a huge number of pictures of star streaks; short exposures would miss any putative KBO. Even if you got a picture with a nearby KBO in it, NH does not have the software capability to compare images and identify images with potential KBOs in them. That means you'd need to download each image to look through all of them, which would require a dedicated 70 meter dish just for NH telemetry.

Other question was about spinning but pivoting 90 degrees and taking short duration LORRI photos when you happened to be looking forward, to build up an image of KBOs ahead of NH. IIRC problems were that put LORRI pointing too close to the sun during the backward facing part of the rotation, and NH does not have the software capability to stack and align images or to compare them, so you're back to needing a 70 meter dish for this to work.

I'm a little confused why the JWST is being mentioned when it isn't scheduled to be launched until 2021.

NH has power until the mid2030s; once JWST is operational, it may/should be sensitive enough to identify any KBOs that cross NH's trajectory from mid2020s through mid 2030s. That could still give NH enough time to maneuver for another close flyby.

[nprev]

If I recall correctly, and can't remember where I saw the reference, Alan said that it's expected that candidate objects will be at V=32 or some other incredibly low value in the 30s as seen from the perspective of all our instruments back here at 1 AU (more or less).

That's gonna be extremely challenging, though hopefully innovation will prevail as it has so often in the past. Still, it seems by far the most likely that NH will find its own next destination via LORRI.

[stevesliva]

How many LORRI photos is "many" when you're just looking at the cone of potential flyby targets? You definitely do not need to look near the sun or to the side to search for flyby targets. NH would take X number of images towards its future path every Y amount of time... X might be small and Y might be large for things to be feasible.

But I will say that the slow-ness of downlink for UT makes me realize that you might be right in thinking at least some processing must be done on board.

Slide 3/4 here seemed to be hinting at what they hope the future holds for LORRI.
http://pluto.jhuapl.edu/News-Center/Press-...page=2019-01-03

[HSchirmer]

How many LORRI photos is "many" when you're just looking at the cone of potential flyby targets? You definitely do not need to look near the sun or to the side to search for flyby targets. NH would take X number of images towards its future path every Y amount of time... X might be small and Y might be large for things to be feasible.

But I will say that the slow-ness of downlink for UT makes me realize that you might be right in thinking at least some processing must be done on board.

Slide 3/4 here seemed to be hinting at what they hope the future holds for LORRI.
http://pluto.jhuapl.edu/News-Center/Press-...page=2019-01-03

Some of the more engineering level detailed discussion starts up around post 300 of the Pluto/KBO > New Horizons thread.


My back of the envelope calculations back then were based on a .5 degree flight cone, and KBO denisity estimates to guesstimate that based on Lorri specifications are 17.5 magnitude 9.9 seconds exposure. stretching the camera to 29.9 second exposures, NH might detect 28th magnitude KBO's at 2/3 of an AU, which could mean tens to hundreds of KBOs would be "visible" as they pass; we just don't know when to turn on the cameras, or which way to point.

[WTW]

I'm a little confused why the JWST is being mentioned when it isn't scheduled to be launched until 2021. https://jwst.nasa.gov/about.html

Yeah, that is a problem -- and assumes there are no more delays.

But NH will just be finishing up transmitting back Ultima Thule data in 2020, and will still be in the "cold classical" region of the Kuiper Belt in 2021, with the possibility of being able to divert for another fairly close encounter.

So IF adequate time can be scheduled on JWST, it has the angular resolution and sensitivity in the near IR to be able to detect potential flyby targets that NH could still get to, that no other telescope can. It just won't be any time soon. :-(

[stevesliva]

... ... ... NH might detect 28th magnitude KBO's at 2/3 of an AU, which could mean tens to hundreds of KBOs would be "visible" as they pass; we just don't know when to turn on the cameras, or which way to point.

I almost didn't follow this segue. I have been stuck on the "see them coming" part of things. But you're right... even if you seem them "coming" how do you look at them as you go "past," or even better, go past them close enough to resolve them. How could you figure out orbit / TCM needed in time? Is that truly feasible? It would be neat if so... but this only shows signal a few weeks out:
https://en.wikipedia.org/wiki/(486958)_2014...oach_video1.gif
... NH would be past UT by the time it saw it coming. Effectively. And see only a pixel.

[HSchirmer]

I almost didn't follow this segue. I have been stuck on the "see them coming" part of things. But you're right... even if you seem them "coming" how do you look at them as you go "past," or even better, go past them close enough to resolve them. How could you figure out orbit / TCM needed in time? Is that truly feasible? It would be neat if so... but this only shows signal a few weeks out:
https://en.wikipedia.org/wiki/(486958)_2014...oach_video1.gif

... NH would be past UT by the time it saw it coming. Effectively. And see only a pixel.

There's the anthropomorphic problem here; NH doesn't "see" anything.
It's only when the pictures are downloaded and analyzed by somebody on Earth, that anything is "seen".

Biggest problems are going to be time lag and fuel consumption
A "look ahead" picture requires that you spin-down the ship, pivot towards the aim point, hold steady while you take a series of 29.9 second exposures, then, pivot back, spin up again and wait for your photos to download at around 1k- bit, about the top speed of that iconic acoustic modem from "Wargames" or the silver "1200" brick modem that came with a Commodore 64.


While it seems that onboard software processing is the way to go, because it leverages the advantage that NH is already out there, that requires somebody to write the software to layer multiple images together or do AI based flip comparisons using a PlayStation 1 CPU.


I suspect it's doable, but you're going to need some impressive software engineering to do it.