2014 MU69 "Ultima Thule" flyby, For discussion of the encounter as it happens Options

[elakdawalla]

It's only 10 days now until the "Core" phase of the 2014 MU69 flyby begins! I thought it was time for a new thread. Carry on discussing distant observations of other KBO worlds in the KBO encounters thread, and use this one for MU69 until after the departure phase is officially over on 8 January.

I'll be posting a "What to Expect" article next week.

[Req]

I tweeted this at Alan Stern but I imagine that he's getting so much spam that I'm going to repeat the question here in the off chance that he gets some time to read it and give a reply:

@AlanStern Amazing work by the whole team. I'm curious why thumbnails of CAO6 weren't prioritized among some of the first data? Couldn't those be binned down to more or less trivial amounts of data by the spacecraft and still give you an idea of what to expect? Thanks!

Since I'm in long form - GIF compression with white and black only, say 4x4 each, then maybe even reduce that to a 4-bit value that simply represents the number of pixels that ended up white(and if you are really trying to conserve bits, then gzip the whole 4-bit series as a C struct/array with the sequence identifiers as the keys, I imagine it would compress to <300 bits or something ridiculous like that) - all 3 of these options are practically nothing even at <1000bps, and the binning process sounds like a good deal less work than the process described during today's press conference to try to infer if and in which sequence there is anything useful based on refining trajectory elements, so I assume that there's a very good reason for not using thumbnails and I'm very interested to learn what it is.

[mcaplinger]

I assume that there's a very good reason for not using thumbnails and I'm very interested to learn what it is.

Not having to have the ground in the loop to decide what to send down would be one reason. Especially when you're six light-hours away.

[Req]

Not having to have the ground in the loop to decide what to send down would be one reason. Especially when you're six light-hours away.

No need for the ground to be in the loop for the operation I described. What I described is a simple sequence of software that the spacecraft can run to produce a data product that could be less than 300 bits which would tell them if, and in exactly which images the highest resolution pass captured UT. It's not unprecedented, this has been a common practice during the opening days and hours of multiple recent missions and encounters(thumbnails in general that is, not the extra reductions I described to further minimize the size of the useful dataset in the case of this particular mission.)

[mcaplinger]

What I described is a simple sequence of software that the spacecraft can run to produce a data product that could be less than 300 bits which would tell them if, and in exactly which images the highest resolution pass captured UT. It's not unprecedented, this has been a common practice during the opening days and hours of multiple recent missions.

Which missions are those?

Maybe they just didn't think of it or they couldn't make it fit in their FSW. There's a general bias to minimize work done on the space side even if work on the ground side is 10x harder.

We use thumbnails on MSL, obviously, but usually not to pick and choose what to send down once we see the thumbnails. Sometimes, not often.

[Req]

Which missions are those?

Maybe they just didn't think of it or they couldn't make it fit in their FSW. There's a general bias to minimize work done on the space side even if work on the ground side is 10x harder.

We use thumbnails on MSL, obviously, but usually not to pick and choose what to send down once we see the thumbnails. Sometimes, not often.

MSL is the one that comes to mind instantly, however I've seen the same practice being used on Cassini for encounters, ROSETTA did it, and I just assume it's being/was used for insight, since phoenix did the same. Perhaps the piece that I'm missing is just your third sentence about 10x ground. Thanks for that insight.

[mcaplinger]

however I've seen the same practice being used on Cassini for encounters...

Did Cassini use thumbnails? Maybe, I wasn't aware it did.

NH probably wasn't designed with a flyby of a small body with such uncertain centering in mind, I don't think they had so much of a problem at Pluto (but I could be mistaken.)

FWIW, I argued long and hard that thumbnails for MSL Mastcam were not worth implementing. In hindsight I was wrong, but the use case of ground-in-the-loop selective transmission doesn't end up getting used that often, and we had to work fairly hard to get the thumbnail products small enough to be cheaply downlinked. They are mostly useful just to show that a sequence executed as designed, but that's fairly expensive for what amounts to debugging messages (though one can amuse oneself by building mosaics of thumbnails while waiting for the full products, etc.)

[WTW]

Did Cassini use thumbnails? Maybe, I wasn't aware it did.

NH probably wasn't designed with a flyby of a small body with such uncertain centering in mind, I don't think they had so much of a problem at Pluto (but I could be mistaken.)

FWIW, I argued long and hard that thumbnails for MSL Mastcam were not worth implementing. In hindsight I was wrong, but the use case of ground-in-the-loop selective transmission doesn't end up getting used that often, and we had to work fairly hard to get the thumbnail products small enough to be cheaply downlinked. They are mostly useful just to show that a sequence executed as designed, but that's fairly expensive for what amounts to debugging messages (though one can amuse oneself by building mosaics of thumbnails while waiting for the full products, etc.)

Emily mentioned that histogram data are included in the image metadata that is being downloaded before the rest of the images are -- which should be enough to determine which frames contain UT image content, and to help prioritize them.