Literally a wiring error.

Quote from the talk:

Thanks for the quote.
Anyway I think attempting a firing is a very very ver bad idea, taking into account MUPUS result: we didn't land on a "dirty snowball", we landed on a "snowy rock",as somebody said.
Firing an harpoon from 50cm toward a rock rather than from 2 meters toward ice/snow would completely destroy Philae by harpoon bouncing back, IMHO.

I got the time to view the video, which is 1.5 hours in duration. I offer the following for those who haven't the time to view it's entirety but might make use of pieces:

1. For the many who have done the excellent simulations the flight of Philae to its final destination, using image processing and various trajectory generators, there is a succinct chart at 56:45 offering a time plot of altitude, radius, and speed for Philae from first contact to final stop. It might be useful to tweak your models. I'm unsure of what to make of the radius chart.

2. The harpoon portion of the talk begins at 45:47, then continues during the Q&A around 1:14:30. Jmknapp quotes the best statement in a post above.

There has to be software in the "wiring" else we couldn't consider firing it now. Not quibbling with wordsmithery, but "if we had fired the right sequence, it would have worked" would have expressed the same facts, and set the basis for a discussion of should we fire them now or not? Can we fire just one? One at a time? How much flexibility is there in the software/hardware lashup?

I support the notion firing at least one harpoon, since it does a lot of in situ contact science, a lander's dominant purpose. However, there are enough topics in the pros and cons of firing one or more that I think I'll start a post on that topic as soon as I can get to it. (MOD NOTE: Please don't. That would be far too narrow a topic. Also see rule 2.8.)

Good suggestion about fitting our simulations to that data, it's doubtful we could generate a better guess at Philae's location but would be great to produce some updated visualizations. The radius data is especially useful since it will pin down a scale and CG for the model (either ESA's or Malmer's). It would also be interesting to see how gravity varies in their simulation by differentiating the radius data twice. It should not be hard to translate the results in that chart into a Cartesian trajectory around the model. I may try and extract the data from that chart image and post a spreadsheet.

MUPUS just confirmed a hardness beyond its limit, somewhere above 2.0-2.5 MPa. The harpoons have been tested to max. 14 MPa. There's a lot of room inbetween that.

I tried that using ImageJ and exported the altitude and radius data to CSV (attached). There's a bit of noise and the speed data was too noisy to use I think.

altitude (Google spreadsheet)
radius (Google spreadsheet)

Click to view attachment

There are a few controllable relays in an e-box which can be switched to fire the harpoons, the relays are controlled by software. So there is the chance to fire the harpoons differently than the first time. Due to the known problems only both harpoon pyrotechnical devices can be fired at once, so it either try both and get bot or try one and be nearly sure that it will burn the wire without firing.

Can anyone tell me the length or radius of the landing legs from the center mounting to the center of the feet? That info seems to be rather well hidden on the Web.

Rob

Philae found ?
http://blogs.esa.int/rosetta/2015/01/30/wh...ill-it-wake-up/

(not sure at all)

It's not found, but they appear to have released one of the search images at full resolution (which makes it rather interesting in its own right). The picture contains the post-landing ellipse, which is great (although without a version lacking the ellipse, Philae could actually have been obscured from our view ).

I guess that OSIRIS pic with what is suspected to be Philae in front of a crater rim has not been released until now, either; so it's a blog entry with lots of interesting new stuff.

The OSIRIS pic is fantastic... But what is this ? (at right of the ellipse) :

There are a couple other ghostly smudges below the red ellipse. I suspect they are splicing artifacts (the image is 2x2).

The new "Philae above the comet" view is pretty subtle. At that position, there could easily have been a little peak/boulder that was catching the sun. Presumably their identification is based on this being a reasonable location given what was already known about it's post-first-bounce trajectory. Or, perhaps there was another frame that didn't show that spot.

Perhaps the shadowed area in the red ellipse of the OSIRIS search image can be used to refine the shape models in that area. If so a 3-D landscape view can be simulated from various trial surface locations and compared with the CIVA mosaic? It looks like we could conceivably get a fit as we will see preferred azimuths with lower and higher horizons.

FWIW, they showed the "Philae above the comet" image at AGU.

The 2x2 image of the landing area is absolutely full of splicing artifacts. With the comet's fast rotation rate, shadows move between frames, so there's going to be some artifacts no matter what. But if you crank up the contrast you will find all kinds of ghost images where the two frames were not totally aligned and then blended with each other. I'm having a hard time getting excited about posting this. I want to wait for the PDS data so I or some of you can do a better job at putting these images together.

IT'S AN OSIRIS IMAGE!

Thank You Holger and team. Some issues, but its full of detail we have not really seen before, the "glassy" appearance of the multitude of rounded boulders below Hetmehit, just like a small region on the mixed material boulder in the CIVAS images. It looks like where ice has melted and refrozen to me. The same as the "goosebumps" perhaps? Lots of pits and hollows in both cliff faces and boulders, large and small. The rough scattered shrapnel we saw in the ROLIS landing images and the large amount of bright highly reflective patches and boulders. Will need more time to find all the little nuggets in this image.

I have posted these images and an extended explanation on the Rosetta blog, but it may not get through moderation until Monday so I shall post the images here without the long explanation. Others will no doubt disagree, but the evidence is strikingly similar to that i found in the Nov 25th NAVCAM image and I remain pretty convinced. No its not conclusive proof, but it may be the best evidence available until bound orbits are possible again in 2016.

https://www.flickr.com/photos/124013840@N06...in/photostream/

https://www.flickr.com/photos/124013840@N06...in/photostream/

These are similar views from November.

https://www.flickr.com/photos/124013840@N06...in/photostream/

https://www.flickr.com/photos/124013840@N06...in/photostream/

HUSH... HUSH... Seems like a little crack eventually appeared on the dam holding back those precious, marvelous OSIRIS images!
The good news seems to be that the close 6km "swoop" is indeed going to happen (next Feb 6th or 14th? Different sources cite different dates),
with the clear objective to take images of the surface of the comet with unprecedented detail/resolution and with a phase-zero angle: no shadows (well, as close to that as possible).
The bad news (Uhm... Ok, let's say I'm a bit biased on spotting where little Philae ended up hibernating) are that no specific attempts will be made to modify the planned orbit trajectory of Rosetta spacecraft to try to get images of the "suspected" final landing zone of the lander. Yes it's a bit on the down side of things but I cannot disagree with such decision. I'm ready to bet my "family jewels" that all attempts so far to "visually" locate the lander haven't failed much due to the "resolution" of the available images (OSIRIS team has them all) and the varied, astounding, nature of the comet's surface. No matter how precise the reconstruction of Philae trajectory is at the moment, there are too many areas, wide areas, of the surface of the comet in the "suspected" landing strip that are just black. And really, really black.

That "swoop" at 6km with zero shadows seems to be the last good attempt.
Go Rosetta!

PS: A bit convoluted. My reasoning is that even the OSIRIS team, albeit in possess of very detailed (for the time being) images, only has images with too many "black" zones to come to a "visual/scientific" conclusion about the final spot where Philae ended up. The up-close 6km flyby becomes therefore the last chance to spot Philae in images... Untill much much later maybe.

My thoughts exactly. A very good estimate of the lander location is given elsewhere in the video and I think it would be hard to improve on without visually spotting it.

I love the flights and watch them religiously so thought to alert the simulators to the chart's existence.

After Philae reawakens, running the CONSERT experiment with an orbit rotated 90 degrees WRT the orbit that produced the current estimate should narrow the long dimensioned error bar down and facilitate the visual search if it hasn't been found.

ADMIN NOTE: I think we can leave the commentary on the OSIRIS team and the recent images releases for now. Everyone has had their say over the last few months and we must all remain mindful of Forum Rule: 2.6. Thanks everyone.

OK. Where can the video be found ? Thanks.

Another point is that the new release contains better density estimates than we've seen before (0.47 g/cm^3), so coupled with a good shape model and assuming uniform density, those doing gravitational simulations should be able to do a better job now.

Actually the fact that it is quite near the equator is pretty irrelevant--it has more to do with the local shape of 67P near that point, and the orientation of that shape relative to 67P's rotation & the sun's position.* Around time of Philae's landing, the sun was above the horizon for ~4 hours per day as seen from a generic spot in the landing ellipse (exact details depend on the exact spot, and even more so on exact local topography) but for the entire 4-hour period the sunlight is coming in from a very oblique angle. In short, the sun is above the horizon for hours but it never gets very *far* above the horizon. So it's very easy to imagine all sorts of local topography that could cut down the amount of sunlight received by any particular spot from 4 hours down to 90 minutes or any other value from 0 hours to 4 hours. That's one reason we haven't gotten good 'non-shadowed' photos of the region--sunlight is always coming in at an oblique angle, thus in a very rough type of terrain, there are always a high percentage of shadowed areas.

*Think of a globe that you sliced in half through the equator and discarded the entire northern hemisphere. Now the entire 'northern equatorial plain' gets the same amount & angle of sunlight that previously only the north pole got, even though the entire northern equatorial plain is 'close' to the equator.

That's all going to change as the seasons do, however. The comet's seasons are long, but not that long. That's why the team is confident enough in a wake-up to already be planning science operations for a 'second phase' of Philae's mission later this year, around perihelion.

We should be able to assign a "cometographic" latitude to the expected Philae location? This is the angle of the local terrain normal to the rotational equator. If we had this I could update the solar tracks (presently assuming we're on the cometographic equator) in my CIVA mosaics.

Along those lines, I released this updated simulation of Philae's ride across the comet, from Philae's point of view a few weeks ago:

https://www.youtube.com/watch?v=agmmOT8N-bk

The *big* update is that I managed to get the location of Philae's final resting site wrong on the previous videos (just plain made a mistake in placing it, because I was using the ESA's 3D model at that time, which is pretty lacking in detail. As soon as I started comparing the illustrations of the Philae final location ellipse to Malmer's much more detailed model with the much more detailed surface textures added, it became obvious that I'd placed Philae's final resting place in the wrong 'canyon' in the crater rim. Aargh.)

Good news is, moving the 3rd/4th impact points to their more realistic locations makes 2nd bounce more of a 'graze' and the distance traveled between 2nd & 3rd bounces more believable. (The direction/distance traveled between 2nd & 3rd bounces was my main point of discomfort with this simulation before--now that issue is rather neatly solved.)

Keep in mind that this isn't a physics simulation per se, it's more of a visualization of Philae's "Wild Ride" across the surface of 67P with known constraints about Philae's location factored in as best possible. With that in mind, here is some assessment of what is real and what is more of a guess:

• Descent & landing spot - based on the ESA SPICE data, but adjusted to end at the known landing spot; this should be pretty close to reality (location is -4 m different from ESA radius data just released)
  
• 15:43 photo - I can duplicate this photo of Philae from Rosetta's location within 30 seconds of 15:43, so we have Philae's position at least pretty close here. (Radius of the location in the video at 15:43 is +26 m from ESA radius data just released.)
  
• 2nd impact - because of the known 1st impact point, 15:43 photo, and known 2nd impact time of 16:20, this location is actually pretty tightly constrained. I'd say 2nd impact is quite certainly within a couple hundred meters of the point shown and maybe even within a few tens of meters. You have some freedom of choice in the the vertical direction, but not so much in the horizontal direction or you just can't get the timings to work out right. At any rate, this impact must have been somewhere along the crater rim in this general area. (Location I've chosen in the video is +27 m compared with ESA radius data at that moment)
  
• 3rd impact - this point is the one that is much more a guess from among many different possibilities. If we 'know' the (approx.) location of the 2nd bounce (reasons given above) and the (approx.) location of the final resting place (ESA's calculations), and that there is one bounce between the two at some certain time, then that narrows down the possibilities quite a lot, but there there are a whole lot of different possibilities for where that in-between bounce could be. I've just picked one of them that looked possibly possible to me, more or less at random (and it looks like the info I had about the time when the 3rd bounce happened was incorrect, as well). (Location I've chosen is +90 m compared with the ESA radius data for 3rd impact point--this one is obviously the 'most wrong' of all the impact points.)
  
• 4th impact - I've just picked a point near the center of ESA's location ellipse, and that looks like it is in the bottom of a dark cleft. So this is the general location where ESA thinks Philae is, but obviously it is not 100% precise. (Location I've chosen is +4 m compared with the ESA radius data for the 4th impact point; Malmer's 3D model doesn't have much detail in this area so don't make too much of that one way or the other.)

FWIW, looking at the recently release radius and altitude data, at first glance I would say the main discrepancy between this & reality will be the 3rd impact point. Instead of hitting left and above the 4th impact point with a final rightward/downward bounce as shown in the video simulation, it looks like the 3rd impact point was more likely right and below the final impact point. Also it was at a different time than I've shown. With the elevation & radius data, you might be able to narrow down the possible trajectories quite a bit.

My first (wrong) simulation had Philae flying across the crater to the rim on the opposite side. It's still a great view of the top of 67P's head and an example of what could-have-been if Philae had hit a piece of the crater rim with just a slightly different angle.

This second simulation shows Philae more 'riding the rim' of the crater from 2nd impact to final resting point.

If my speculations just above are correct, Philae must have been even more 'over the rim' for the 2nd-3rd impact time. Maybe something about half way or one-third of the way between my 2nd simulation & Malmer's simulation.

A few final observations:

• Per the very nice graphs posted here, trajectory after 3rd impact looks very different from and smaller than trajectory after 1st impact and 2nd impact. My guess: Philae hit much more on its feet at 3rd impact than 2nd, thus absorbing more of its velocity--the vertical component (which we can see from the newly released ESA data) but also (I'm guessing) this impact ate much of the horizontal component Philae's velocity.
  
• One reason Philae stopped so neatly on impact 4 is it was near the peak of its vertical movement at that moment, ie, by good luck, the vertical component of its velocity was minimized/near zero at that moment. (Again, see graph here.)
  
• Malmer's very early simulation of Philae's flight is looking more and more prescient. Malmer didn't know about the 2nd impact at that time. That 2nd impact graze knocked Philae a bit to the left and a bit upward. ESA's final resting place for Philae is just a little (couple hundred meters?) leftward & upward from the point Malmer identified.
  
• At 3rd impact, Philae actually had about the same or maybe even higher speed/kinetic energy than at 2nd impact (3rd impact point is 40m lower than 2nd impact point, so if if the 2nd impact didn't absorb much energy, it may have even had higher speed at 3rd impact than 2nd impact).
  
  So . . . it was purely a matter of chance and the precise configuration of the 3rd impact point, that Philae took a shorter, lower bounce and came to a stop. It could just as easily have bounced in a different direction and traveled just as far (or further?) from 3rd bounce to 4th bounce as it did from 2nd to 3rd. With more of a grazing collision again, it could have ended up more in the neck area, or who knows where!
  
  Even if it had just bounced to a slightly different position, it could have been in a place entirely lacking in sunlight.

Hi Flug. Its getting closer, knowing that 2nd touchdown point would make things a whole lot easier.

http://www.unmannedspaceflight.com/index.p...st&id=34906

From this data at about 83 minutes from first touchdown, Philae started to increase in speed AND altitude, with the radius still decreasing. Would this not indicate the point where Philae traveled beyond the rim, a point where the surface starts to fall away AND the influence of the higher gravity of the neck region and the comet's rotation become factors to increase Philae's speed? This would mean the third impact (2nd touchdown) was below the level of the crater rim as you suggest. This might explain the +90m difference in your radius figure.

Back in early December after the ROMAP data first emerged, without any calculations or numbers, this image was an attempt to visualise for myself, Philae's journey. The spur of comet that I suggest is the final resting place for Philae, sticks out some way from the rim of the crater, how that affects the radius figure I am not sure. Ignore the pink line in the first image, that just tracks the line of the OSIRIS images as Philae was approaching the first touchdown.

https://www.flickr.com/photos/124013840@N06...in/photostream/

The You Tube video flightpath noted in this second image is Matthias Malmer's simulation mentioned in your post, (at the time I did not know it was Matthias's work).

https://www.flickr.com/photos/124013840@N06...in/photostream/

Philae appears to have bounced UP into her perch in the crevice, rather than down, as I assumed in my images. Don't know if these ideas help at all, but I for one appreciate your efforts and would be very interested to know if a 2nd and 3rd touchdown on that spur fits any better.

P.S. Wasn't the total flight time 118 minutes, 15:34 to 17:32?

Thanks, that is very helpful - original data source here at 56:44, just in case anyone missed it upthread.

I made this little addendum to your sheets, which uses the data from those two charts to calculate and graph the altitude of the ground under Philae:

Ground height under Philae - google spreadsheet

Ground altitude is just a simple subtraction between the charts you give--if you know Philae's distance from the 67P, and the altitude of Philae above the ground, you can subtract to find the distance of the ground from the center of 67P. Note that times of the radius/altitude graphs don't *quite* line up but they are close enough that I haven't bothered to try to resolve the relatively few seconds' discrepancy.

I made a quick annotation of that graph, attached.

If this data represents Philae's real flight, it basically confirms the general concept of where we think Philae went after 1st touchdown--over two small craters, then 2nd impact on the side of "Mt. Malmer" (the ground profile plus altitude of impact point given in these tables narrows down the potential impact points to one spot on the ridge on the side of "Mt. Malmer"). Then it goes over the large crater (Hatmehit) rim.

The data (if it truly represents Philae's position at these times) is pretty helpful in nailing down Philae's position up through the point it goes over the Hatmehit rim. But after that our knowledge of 67P's exact topography is pretty vague (still not many good photos of that area; Malmer's 3D shape definitely isn't precise in that area) so knowing exact altitude is helpful but maybe not as helpful in defining a very precise location as you might hope.

FWIW I'm not quite sure what to make of this data. What is its source? What instruments would give the distance of Philae from 67P center and its distance above the ground (seemingly taken exactly along the radius line to the center . . . )? And the data seems very, very smooth--like actual radar data would never be nearly as smooth. And the ground, for instance, up and over the large crater rim can't possible be that smooth--this data shows the surface smooth to the sub-meter level. In real data, say a radar altimeter, you'd expect to see both rough ground and then some instrument noise on top of that--wouldn't you? And the times of 3rd & 4th impact don't agree AT ALL with those given on the Esa site previously, ie here--the times aren't just a little off, but **many** minutes different.

I'm guessing this data must be from a Philae simulated trajectory or it's been fairly highly smoothed/manipulated in some way.

FWIW the most remarkable thing to me in this data is the very, very low altitude of Philae after Impact 2. It spends most of the time after impact 2 less than 30 meters from the surface, and lots of time--for instance, the whole jump over the crater rim, at less than 15 meters altitude.

If this data is an accurate depiction of what Philae did, it was quite literally skimming the surface as it went up, over, and back down a pretty big mountain in between impact 2 and impact 3.

I've tried to find philae and this is my best candidate to date.
It's difficult to identify features in the left part of terrain with actual CIVA images.
Accurate enough 3D terrain model or stereo image would be helpful for this matter.

Looking at the video now, I see that you pulled that data from a graph in the presentation (graph here in the video). So that explains some of the smoothness. However, I do wonder whether they worked out the Philae trajectory using whatever method(s) they have available, then plotted the result on their 67P model (shown in the presentation) using whatever method(s) they have available.

Once they have done that, it would be very simple for them to print off a graph of Philae's distance from comet center & altitude above the ground based on that trajectory and that 3D model, for illustrative purposes. And then for us (or, more precisely, me) to jump to the conclusion that they have some super-secret source of precise data about Philae's location and altitude that they haven't shared with us until just now.

But if this line of thought it right, it means that these charts represent data from their 3D model and presumed Philae trajectory, and not actual 1st-hand data from Philae's instruments that directly measured Philae's position and altitude above the ground during its flight.

I think that is at least a possibility for what this data represents.

FWIW yesterday I put together a new STK visualization of Philae's flight based on this altitude and radius data--I'll upload it to Youtube soon. Looking at the diagrams in the SETI video now, my current hypothesis for Philae's trajectory looks remarkably like his. (Of course it should--it's based on the same data! But mine was reconstructed from just the altitude & radius data plus Malmer's 3D shape model of 67P in STK. So it gives us at least some confidence that our 3D model etc isn't somehow disastrously off from their data.)

Are there in fact any such instruments? Other than camera-based photogrammetry I couldn't find any evidence that such instrumentation existed in the Philae fact sheets. I don't think the sounding instruments can be used for altimetry.

OK, here are updated videos showing Philae's trajectory from some time before 1st impact until final impact. They again use Malmer's wonderfull 3D model of 67P and have benefitted from feedback & ideas from many of you on this forum. (Mistakes, of course, I can make all by myself.)

Overflight/overview of trajectory:

http://youtu.be/VaMrt1FBSBo

Philae's first-person viewpoint of the flight:

http://youtu.be/5iT3bLm5v10

As I mentioned before, these visualizations have been updated to match the data as shown in the SETI presentation linked upthread, and the trajectory shown here matches quite closely with the trajectory shown in that presentation. I would say it is pretty darn good ™ up to the point Philae crosses the crater rim but gets more vague and speculative after that. Neither we nor ESA know exactly the position of the 3rd & 4th impacts, and also our shape model of the dark area over the rim is pretty vague. So that part of the video is great for giving a general overview visualization for how the flight must have gone, but don't expect it to be super-specifically correct on details.

The most remarkable part to me, if the data from the SETI talk is correct, is how low and neatly Philae flew up and over the rim mountains. Altitude is less than 15 meters for an extended period as it hops over the top of that mountain.

Also note that the horizontal component of the trajectory is pretty close to a straight line now. Just slight leftwards nudges at both Impact Points 2 and 3. As mentioned in the SETI talk, those leftward nudges were a godsend, because if Philae had gone straight, or been nudged rightwards, it would have gone right straight into the dark/winter side of the comet. As it is, it's just on the border of the light/dark areas.

One unfortunate thing: The first version of this video (topview, Philae view), with Impacts 3/4 in the wrong location, was still a fine visual tour of the head of 67P, taking you past some really amazing cliffs, boulder fields, craters, etc.

Version 2 (Philae view), with my first guess at Impact Point #3, was still fine visually, with an interesting 'ride the rim' view of the crater.

But now Version 3 (topview, Philae's view), clearly the closest to the 'real thing', takes you into the Dark Side of the comet for a good half of the ride. Once you've crossed the rim mountains, it's just dark and dank, with not much to see. So as you're watching the dark screen, just remember: Philae really did go over to the Dark Side, and that's what Philae really saw there--not much of anything! It's really, really dark. Probably darker in real life than what is shown in the simulation.


Compared with that previous versions of the visualization, this version:

- Moved Impact Point 2 about 25 meters down and left (now it is right on the ridge of Mt. Malmer, which is in accord with ESA data)

- Reduced the maximum height of the trajectory ellipse from Impact Points 1 to 2 by about 20 meters (in accordance with ESA data; the visualization now matches the 15:43 photograph and data from the SETI talk within a couple of meters)

- Moved Impact Point 3 a fair distance to match the info from the SETI talk (Impact point 3 was always the point we were guessing at the most). This puts far more of Philae's trajectory 'behind the rim' and in the dark area of the comet

- Lowered the trajectory Impact point 2-3 quite a bit--now it just scrapes over the top of the rim, by less than 15 meters at some points. This is in accord with data from the SETI talk.

- Moved Impact Point 4 just a couple of meters to the altitude indicated by data in the SETI talk. This is still not precisely located and could be anywhere within the area ESA indicates in the diagrams linked upthread. (Or elsewhere if ESA is wrong!) Also note that the shape data in this area of the 3D model is vague/incorrect (because it is on the edge of the comet's dark side, there are not enough photos to get good shape data).

- The one piece that doesn't quite harmonize is the recently released photo of Philae just above the rim at 17:18. There is conflicting information from the SETI talk, previously released impact times for Impacts 3/4, our vague 67P shape model for the region 'beyond the rim' where Philae landed, and this photo (assuming it does show Philae). It's impossible to harmonize all those conflicting sources and make them all work. Just for example, the SETI talk has Impact 3 at 17:16 and Philae at 11 meters above the surface at 17:18. Philae can't be both there AND well above the rim (as shown in the photo from Rosetta's viewpoint; clearly it's more than 11 meters in altitude) at 17:18. So the end of the video is more approximately what happened in approximately the right area of 67P, than exactly what happened in exact places.

Remarkable videos with the update. It makes sense to hit the rim (impact point 2) at a lower location simply from considering the odds it would have hit so close to the top. It follows the theme flug mentioned a long time ago about how close it was to going way into the dark side. I would guess the recent OSIRIS image can help with the shape model in the dark areas?

Just for fun I put together a comparison of ESA photo of Philae at 17:18 vs the STK simulation of Philae's position at 17:18.

ESA's photo shows Philae highlighted by red crosshairs, the STK simulation shows Philae as a blue dot highlighted by a red circle.

It really shows how amazingly awesomely realistic Malmer's 3D model really is--it's like having your very own comet in your front room, in so many ways--but also the discrepancy between model & reality in those areas where there wasn't much data when the model was made. It's just a good little reminder that we tend the trust the 3D model because it looks so good; it's worth reminding ourselves that it has limitations, too.

For example, the exact point on the rim where Philae is, in real life is sloping upwards quite sharply, whereas our model has the same area as a more flat-ish pass through the mountain.

Regardless, you can see that Philae is pretty much in the right spot at the right time. Obviously, if this photo does show Philae at 17:18 on Nov 12th, the ESA (and our simulated trajectories) are on the right track in general terms.

FYI the STK simulation/trajectory used here is similar to my Version 2, where Impact 3 is more up on top and then veers a bit right to Impact 4. The SETI data (my Version 3) has Philae so far below the rim at 17:18 I'm can't image how Rosetta could have had it in line-of-sight. There might just be something about the geometry that I'm missing, though.

Keep in mind, no one is sure if the ESA 17:18 photograph really shows Philae or not. Might be another rock, or just a random glint. It's very, very suspicious just because it's so close to where they think Philae ended up just a few minutes later.

Flip side: If that really is Philae, just take a look at how hard that shape/color is going to be to pick out of the surrounding terrain, once it has terrain behind it. It's like nearly perfect camouflage.

Download full-sized images here: STK simulator image - ESA real image - Animated comparison

Excellent update Flug. I wondered where that "smooth" altitude and speed data came from. Your surmise seems spot on to me, its taken from their flight simulations. Remember the third refinement of the CONSERT search area also had input from the Flight Dynamics team to improve it, possibly this data.

My skills and available tools are way more limited, but I have also updated my 2D guesstimate or "artist's impression". No calculations or physics, done by eye and dead reckoning from the images. I should have watched your video first as I have impact point 2 on the crater rim the other side of the valley from mount Malmer. If that WAC image is correct and that is Phillae, she must have dived down the side of that mountain, yes their speed data suggests an increase in speed, but to hit the surface from there in 7 minutes?. Anyway here are the three images I have come up with. Note I have marked Matthias's "Glint", a large, fractured, exposed icy boulder, only a few tens of metres from my pick for Phillae bijou residence.

https://www.flickr.com/photos/124013840@N06...in/photostream/

https://www.flickr.com/photos/124013840@N06...in/photostream/

https://www.flickr.com/photos/124013840@N06...in/photostream/

Pretty similar conclusions, it seems given the inconsistencies in the times and data, a basic visual approximation serves as well for the time being. I do like the ride over comet a lot better though, thanks. :-)

The reconstructed trajectory is a simulation with each clue from the various real data sources used to constrain the results. One or two slides after the chart of data they mention how they can use simulation to infer information about the soil mechanics attributes and a software package called SIMPACK is listed... so that may be the tool used for trajectory simulation (I've seen this used in Philae landing gear analysis too, here). Definitely too smooth to be real data anyway.

The new visualizations are beautiful, especially the first-person view -- great job!

I'm so happy to see my model coming to good use. Will try to update it to include better coverage of the landing area. I have all the data I need. It just needs a little more work...

Feels like Xmas when Malmer releases a new model!

Do you have a textured version? If I remember correctly, you build the model from a type of bundle adjustment on the camera poses to minimize reprojection error for interest point features, right?

I should do a new textured version.

It is slightly tricky to generate texture-coordinates and texturing it in a semi-automated way.
I want to minimize stretching in the texture so a standard Lat long projection is out of the question. I will probably break the surface into regions and use some simple mapping for each. (so that I can update the mesh without having to redo everything.)

The last release was just manual work all the way trough.

Watch this space.

Latest report from ESA seems to indicate they won't be doing a flyby to look for Philae, but going ahead with the close flyby at 6 km with the Sun directly behind them to get shadow free pics, as originally planned. Presumably they are going to wait to see if and when Philae will wake up, then try spot him.

I think I have found an image of Philae in the ESA_Rosetta_OSIRIS-NAC_Landing_site_montage which would provide a different trajectory then the one assumed to date. I used the Philae_s_primary_landing_site_from_30_km image as a baseline for this area since the Rosetta_OSIRIS_NAC_Philae_touchdown_mosaic blocks it. (note the preview doesn't loop the animated gif probably have to download it)



To cross check it I took the key 15:43 picture of Philea at the same scale and removed the glint, I also provided a comparison from the recently released Philae_descends_to_the_comet gif



Also see the Philae_descends_to_the_comet gif to demonstrate that glint orientations are actually not so common and the 15:43 shot was pure luck. This dark image of Philae is what is more expected

Click to view attachment Link 2.4mb
ADMIN EDIT: Inline image replaced with thumbnail

When I get a chance I'll clean up my image of the trajectory but the ground track roughly appears to go roughly 45 degrees counter clockwise in the ESA_Rosetta_OSIRIS-NAC_Landing_site_montage from the current ground track. It also seems to match well in several places with the blurry CIVA image at landing plus 5min? - which I took a stab at cleaning up as well - )

To add to my last post - I created a image of a few possible Philae ground track trajectories. What's missing is the exact time of the ESA_Rosetta_OSIRIS-NAC_Landing_site_montage - it's anywhere from Landing +2 to L+6min - I've assumed about L+3 since the picture with Philae is at L_8.9 min and based on the blob and shadow changes



Here's a high res link: Philae_primary_landing_site_possible_groundtrack.jpg Full Resolution

I think ESA's reconstruction should be close to the true trajectory and so I'm not convinced that is Philae. However, I do think that might be its shadow since the sun vector lines up nicely.

I can "see" at least 5 Philae in this image. One is even holding MUPUS. (But it is in too much lighty place to be actually it).

How many pixels is Philae supposed to be large in this image? 2? 20?

From the blog post:

Brian, as to your Philae shadow hypothesis - it may be so, but I think it is the actual lander for a few reasons.

1) The deck of the object is variable colored - inconsistent with shadows (see figure of some expected shadows at various heights) (The right object is an earlier Philae photo so remove a leg or two - due to the view orientation - but it illustrates the concept.)
Click to view attachment

2) The ESA assumed trajectory doesn’t fit with this being a shadow because it doesn’t line up with the Nav cam image shadow line region (2 dashed blue lines) and there is no shadow in the plain where it would be expected at 15:43 if you extrapolate the hypothesized shadow (blue line). And if in fact the height is only 66 meters then the shadow should be visible in the 15:43 image - there is nothing in a bare area where a shadow is roughly expected. There is no shadow along that entire line - except in the 15:43 image we have a annotation box blocking the very close shadow line - but if there were a clear shadow there I’m sure ESA would have pointed that out in the mosaic they made. see below image (no thumbnail) which has a sun line drawn on the 15:43 image.

OSIRIS_spots_Philae_drifting_across_the_comet-no-shadow for current ESA trajectory if object is Philae's shadow

Also I have taken my image and annotated the current ESA reconstructed trajectory (dashed lines). Red is ground track, purple is observed image and light blue is shadow track assuming the Navcam was a valid shadow image

Click to view attachment

My issue is we have an huge clear object with many resolved pixels - it is too big to be a rock and is exactly Philae sized and colored that is at a disagreement from the reconstruction to date. I think it merits further consideration. I would note that the ESA reconstruction is based on many assumptions and very talented educated hypothesises - but I think the only piece of data that is iron-clad is the 15:43 image. I think this is another piece of iron-clad data.

I have also better aligned the contrast of the two images for a cleaner comparison of features (download to get the gif to animate)

Click to view attachment

You are putting a lot of work into the analysis and I don't want to sound like a naysayer but keep in mind that the shadows will not be following a straight line!

Assuming Philae is following a straight trajectory is close enough to be valid (at least for the duration we're examining) but the terrain is undulating enough that the shadow will definitely meander around and so there is no way you can conclude it will be along one of those lines.

Brian,

of course you are correct - predicting shadows on such a variable surface is very tough without a detailed simulation and elevation model, and I have made a simplistic linear extrapolation (though probably not that far off for a short time/distance). However the sun-line from the expected Philae position at 15:43 (assuming the ESA reconstruction is valid) is completely void of any shadow candidates and the relatively flat plain where the object shadow line would be extended (solid blue line) and intersecting with the 15:43 sun-line has nothing in a huge uncertainty range. It's not so much as saying that the shadow can't be there, as saying *if* the object was a shadow (which I don't think it is) - and the plains are relatively flat for this relatively short distance/time duration - it is highly likely that a subsequent shadow would also be visible in that area with an uncertainty region around it. There is absolutely nothing there or around it, which is not proof by any means, but adds to the evidence that it is more likely that the object is in fact a direct image of Philae. (I wish ESA would give us their expected shadow location in the 2 images given their reconstruction - not to mention exact image times!).

If I get some time - I'll try and gin up flug's sim and see if I can get the 15:43 image and this likely image of Philea to fit a consistent trajectory (or family of them) - (I wish I could convince flug to do it! as Philae might wake up first before I get to it!)

Hey,
i'm new here.

i also started seaching on the pictures provided by esa. I tried to extrapolat the direction of the probe.

http://imgur.com/DHCFgpe

It looks very much like the missing probe to me, however the position is far of the position marked by esa. It would be interesting to know when the shoot of this photo was made. If it was made while rosetta should observe the lander, it seems the probe was still flying.

The lander will be 3 pixels across in that image.
So, no. Not the lander.

(hint: One of the things you're running up against is Occam's razor. If you think you found the lander, set aside your knowledge that the lander exists at all and consider whether the thing you found can be easily explained by it being a natural surface feature.)

Yes i expected this kind of answer, and i thought that thing there is to big. On the other hand i compared the size of this stones, which are placed on that flat area.

These stones on the flat on this picture
http://imgur.com/9ICqv4p

look at least twice as big as on this picture
Click to view attachment

which is next to the picture from esa: http://www.esa.int/spaceinimages/Images/20...cross_the_comet

If Philae is 4 pixel in the last picture, it migth be perhaps 8 pixel in the pixel I posted first.

Ok, i agree it probable is the top of a rock, but philae must be bigger than 3 pixels

Keep in mind that the eye sees what it wants to see! As pointed out by hungry4info, you have to second-guess everything you postulate. The comet nucleus landscape is likely the strangest surface that has ever been explored and a lot of our Earth-based intuitions about ballistic trajectories, shadows, horizons, etc., are inapplicable here. Take a look at photos of spacecraft on Mars or the moon from orbit and you will get an idea of just how different the objects look when confined to only a few pixels... and most of those photos have a lot better lighting conditions.

Another thing we have to remember is that the images we are analyzing are not taken while pointing directly down at the surface, but instead at an angle. Many people have posted images projected onto each other (the CIVA image during bouncing, for example) but with only some zoom and rotation. This is incorrect since any image projection will have some perspective and skew as well. On top of that, the relationship between an object's actual size and its apparent pixel size in an image is quite a bit more complicated than you might imagine (check this out for a brief look at what's involved).