9

The object 2014 MU69 is much smaller and dimmer than Pluto. Even as New Horizons approaches, it will be much harder for its 8.2 inch aperture, cooled CCD long range imager LORRI to get a fix on it until it is quite close.

EDIT:

New Horizons made it's first sighting of Pluto nine years before it reached Pluto. MU69 is going to be about 100,000 times dimmer, which means it will make its first sighting more like nine days before it gets there (roughly) (actually it could be a few months, see this). Thus any navigation using the spacecraft's imaging system may have a narrow window of opportunity to be useful. This is why I think this question is particularly interesting - is there a plan to use the spacecraft's imaging data in any way to guide the flyby, or will it be done only with astrometry of MU69 from Earth's Neighborhood and radar range/rate of New Horizons plus data from its last known address, Pluto/Charon?

Pluto's orbit has been monitored for almost a century, while that of 2014 MU69 only about three years, although with three years of diffraction-limited precision astrometry from Hubble that can go a long way.

But going from spots on an image to a true three dimensional orbit and six dimensional state vectors is not straightforward. It seems there could be a larger envelope in distance/eccentricity that could generate the same apparent motion.

Question: How can 2014 MU69's orbit be know well enough for a close flyby by New Horizons?

I'm also wondering how well has 2014 MU69's actual orbit been plotted, and how close is the flyby expected to be to the target? Are they just aiming for zero impact parameter (straight-on) and letting the uncertainty determine the distance, or are they aiming a little to one side? But these are secondary and I will ask separately if needed. 

                       Pluto       2014 MU69
                      -------      ---------
Diameter              2380 km       ~35 km
Albedo               ~0.58          ~0.04-0.14?
distance to sun       5.0E+09 km     6.5E+09 km
Vis mag from Earth     ~15          ~ 27
Years of observation    87 yr       ~ 3 yr

At the time of flyby:

  • relative velocity ~ 50,000 kph
  • distance to earth ~ 6,650,000,000 km
  • round trip radio ~ 12.3 hours

2014 MU69: http://www.minorplanetcenter.net/db_search/show_object?object_id=486958

uhoh
  • 148,791
  • 53
  • 476
  • 1,473
  • The question How close to 2014 MU₆₉ can New Horizons get? is related, and there are suggestions there that the distance of closest approach could be a few thousand km. I can't imagine that would be possible based only on Earth-based or LEO-based observations alone. – uhoh May 09 '17 at 11:42
  • The minor planet center gives it a U score of 2, which from my back-of-the-envelope means at its range the predicted position might be off by 10,000 km over a decade. That seems good enough to a arrange the flyby. The object is small and dim, but Hubble astrometry must have had enough accuracy--parallax would be available as well--to get the orbital parameters that precise. – antlersoft May 09 '17 at 19:28
  • @antlersoft What would be the present uncertainty in MU69's position, today, without any extrapolation into the future? It's not zero, right? Runoff is a one-dimensional error in-orbit and does not explicitly contain range or out-of-plane errors. Also, I can't do it today but I'll probably ask a separate question about uncertainty and scoring. I'll leave you a message here when I do. Thanks! (note to self: http://www.minorplanetcenter.net/iau/info/UValue.html) – uhoh May 09 '17 at 19:38
  • 1
    Note: MU69's orbital speed is only ~4.47 km/s. New Horizons moves at 14.3 km/s. Course corrections to get on encounter trajectory are not nearly as difficult with these parameters, even with incomplete data. – SF. May 10 '17 at 14:32
  • @SF. just mentioning a relative velocity is not a way to evaluate the error in the distance of closest approach. You need at least the difference in two state vectors at a given time; that's at least six independent numbers, not one. – uhoh May 10 '17 at 15:05
  • @uhoh: this gives a direct worst-case delta-V for making a flyby with both positions though, and with an estimate of time to flyby since target is visible, it gives a good estimate of delta-V vs distance by which the flyby can be improved. – SF. May 10 '17 at 15:23
  • @SF. this question is about navigation, not propulsion: "...can 2014 MU69's orbit be know well enough..." If you don't have the state vectors far enough in advance, propulsion alone won't make up for the lack of information (better wording). – uhoh May 10 '17 at 15:29
  • If you know the state vectors 9 days in advance, propulsion decides by how much you can correct the trajectory until nearest approach time. If your encounter is within 100,000km, with maneuver of 1.5km/s 8 days in advance you can crash the probe into the target. – SF. May 10 '17 at 15:33
  • @SF. can you write that as an answer instead of all these comments? Remember, just seeing a faint "dot" nine days out is not the same as having a state vector. You don't just "aim" for it, you need a distance and a closing rate. I'm asking how this will actually be done in this particular case. – uhoh May 10 '17 at 15:42
  • https://en.wikipedia.org/wiki/Gauss%27_method – pericynthion May 11 '17 at 04:29
  • @pericynthion from your link: "NOTE: Gauss' method is a preliminary orbit determination, with emphasis on preliminary." It's the "... well enough..." part that I've focused on here, this particular case. Orbit determination can only be as good as the uncertainty in the measurements used. – uhoh May 11 '17 at 05:01
  • @uhoh: That's it: you don't need distance and closing rate, just 2 out of 3 components of position and speed. 2 images will provide first axis of the two. Perform a braking maneuver along this axis and take third image, to obtain second axis. Perform another maneuver that reduces relative speed in these two axis to zero. Third axis remains unconstrained and flyby speed and time remain unknown, but WILL happen. Once brightness changes become meaningful, relative speed becomes measurable. – SF. May 11 '17 at 06:59
  • Old KSP method of eyeballing encounters, from times when kerbals had no navball or maneuver nodes: point at the target, wait a while, see which way the target moved over your screen. Burn along that axis, see how that affected target's speed of movement across your screen. Continue burning until target ceases to wander over your screen. As long as curvature of orbits doesn't spoil your trajectory too much, you have an encounter. – SF. May 11 '17 at 07:06
  • 2
    JPL expects to be able to image MU69 using LORRI about 3 months in advance: https://www.jpl.nasa.gov/spaceimages/details.php?id=PIA21588 – Hobbes May 11 '17 at 08:03
  • @SF. that may be true for straight lines in empty space without sources of gravity, or two ships moving at constant speeds on the ocean, but not necessarily true for an ellipse and a hyperbola in 3D. If you can either link to or show the math as it applies to this specific example, that would be great! – uhoh May 11 '17 at 08:23
  • @Hobbes excellent - I've made a note in the edit, thanks! – uhoh May 11 '17 at 08:28
  • 1
    @uhoh but considering the short timescale and the large orbital radius the aproximation of moving in a straight line is not to far of – lijat Aug 19 '17 at 16:21
  • @SF. Why doesn't New Horizons have a tracking function could benefit from a discussion like this. I think it's a good question (especially first question by a new user) but doesn't have any real activity yet (and only one vote). – uhoh Jan 02 '19 at 04:16

2 Answers2

5

The long orbital calculations of Pluto's orbit were done primarily to save fuel, and secondarily to target a very narrow region. The mission intended to fly between the Sun and Pluto and Charon (And Earth and Pluto and Charon as well), which could only be done with very careful orbital knowledge.

There is quite a large amount of fuel reserved for the 2014 MU69 mission. They are still trying to determine how closely they will approach it, but it will be between 3000 km and 20000 km. While I don't doubt they would love to be between the Sun and Earth and this object as before, it isn't as high of a priority.

Hubble will continue to make observations, as will ground based telescopes, to help refine the mission to come.

Also, New Horizons will use it's camera to image it and refine the trajectory, as it did with Pluto.

Aiming straight on would be a bad thing, although for this early in the planning, it would be acceptable. I bet they are aiming straight for it now, but will adjust to one side when the date is closer (No source, but it makes sense) Aiming straight at it would mean that you wouldn't know what side the object actually is, making taking pictures very difficult!

EDIT: I actually did a video on this not remembering that I had answered this here. The short is, lots of observations with Hubble, using stellar occulations, using databases of stars to take in to account the stellar motion that each star makes. Adding in a few observations from New Horizons itself, and we have a trajectory good enough for a flyby!

PearsonArtPhoto
  • 121,132
  • 22
  • 347
  • 614
  • Do you think they will choose a number between 3,000 and 20,000km, or is that the 3 sigma uncertainty? I'm really after the method(s) that will be used to know MU69's orbit (an New Horizon's as well) accurately enough to target a few thousand km's at such an incredible distance and relative velocities (about 50,000kph and 6,650,000,000km) – uhoh May 09 '17 at 11:59
  • 1
    It depends on the ultimate uncertainty they have, which I don't think they know right now. I suspect they will have an uncertainty of maybe 1000 km or so in the end, and most of the ultimate distance will be due to science objectives. – PearsonArtPhoto May 09 '17 at 12:01
  • ...and a round-trip light time of 12.3 hours. – uhoh May 09 '17 at 12:08
  • That really won't influence where it will end up at all, aside from making it difficult to plan... – PearsonArtPhoto May 09 '17 at 12:09
  • My question is about how difficult it is to plan. And round-trip light time limits what kind of last-day or last-hour corrections are possible. The object is much dimmer than Pluto, so one may not get images a half-year in advance like before. M69 may be 100,000 times fainter than Pluto, that's why I've included the albedo in the table. – uhoh May 09 '17 at 12:18
  • The plans for flybys are usually planned well in advance, so... Last minute corrections just aren't done. – PearsonArtPhoto May 09 '17 at 12:21
  • OK, that sort-of sounds like a sweeping generalization that may not really apply to a flyby of an extremely dim 30 km diameter object that is 6.5 billion kilometers from the Sun. I'm not sure this is a case where "usually" applies. Of course "well in advance" is a fairly flexible term too. – uhoh May 09 '17 at 12:26
  • It applied to the Jupiter flybys, and every one that I am familiar with. It usually applies to orbital missions too. The only exception I've seen is systems using AI to improve position or images. – PearsonArtPhoto May 09 '17 at 12:34
  • This question is specifically about an extremely tiny, dim and distant object, an analogy to a Jupiter flyby may not be so relevant. – uhoh May 09 '17 at 13:35
  • GREAT JOB! This really answers in amazing depth, wow! I never expected a video answer :-) fantastic! – uhoh Jan 01 '19 at 14:44
5

Okay, collecting my comments into an answer:

As soon as you have a visual, and have enough delta-V, you CAN get a fly-by, although precise time and speed of flyby will remain unknown until much later (when changes of size of the object in images become significant).

This works only for when curvature of orbits is not significant - gravitational influence of other bodies is minor and trajectories on the distances involved can be approximated with straight lines. It will work for moderate orbital curvature too, if you repeat the process iteratively, correcting the error whenever curvature introduces it. It's definitely not viable if the two bodies are, say, 90 degrees apart in orbit of a planet though. For New Horizons it's viable though - the ~1-10mln km at which it's expected to spot 2014 MU69 is nothing compared to distance from the sun, and ~10 days is nothing compared to 2014 MU69's orbital period.

Take a set of coordinates, with Z axis going craft-target; fix it relative to "distant stars" (so it doesn't change as the objects involved move).

Orient craft along that Z axis.

Take two images known time apart. The target will have wandered over the field of vision a certain distance.

Perform an arbitrarily chosen delta-V burn perpendicular to the Z axis in the direction of the apparent travel of the target (XY plane), and after known time take another photo.

Using simple proportions between times and positions you can calculate the value of the burn needed to get a fly-by trajectory - one that brings the relative speed of the target in the XY plane to zero. Since it doesn't move "sideways" and you move towards it, you will get there eventually.

You still don't know when the fly-by will happen, and at what speed, but all you need is to wait - or keep repeating the process to reduce the fly-by distance further.

Of course if the initial visual contact is way "off", you'll lack delta-V to perform the correction. Still, in case of New Horizons, about 100,000 km of initial error should be enough to get a very close fly-by, and anything above that will just add to the fly-by distance.

SF.
  • 54,970
  • 12
  • 174
  • 343
  • 1
    In practice though, the time of the flyby was known to a second or two (necessary for camera pointing) implying that the distance to 2014MU69 was known to 20 km or so. Does anyone know how this was determined? – Steve Linton Jan 02 '19 at 17:07