Can't get a much more precise position fix, in terms of synchronizing rover position to ground imagery, than identifying a pixel (or several) in that imagery that is known to be the rover at a given time.

I'm astounded and delighted with the progress that Oppy is making toward Victoria crater.

From the overhead satellite pictures,I assumed that the Victoria ejecta blanket would be full of loose sand and imposing dunes. The NASA team indicated that this journey would take a month but as of this moment it appears that the arrival could be much sooner.

Also, I wonder if the unexpected reboot of the rover at Beagle crater caused the sudden rush to get to Victoria crater.

I would also like to thank Tesheiner and the rest of the mapper makers very much for your outstanding work and you guys have really made this rover experience much more enjoyable. Thanks.

Tesheiner, Delta is visible in the latest navcams - in principle the angular separation between Delta and Epsilon would help fix our location. (Technically it would confine us to a curve - I'm not sure of the shape of the curve, but it wouldn't be hard to work out.) In practice they're both not very well defined...

Gamma crater would help even more - I've looked for Gamma but can't see it in navcam.

I've seen that when the navcams were downlinked --actually, almost all features commonly seen only on the pancams are currently visible on the navcams too-- so I made a polar projection of that navcam mosaic to use on the route map (I usually do it) and see how it fits when compared with the basemap.

The result is that there are some discrepancies; the problem is that given the poor resolution and distortion on the navcams I can't tell which sol 914 position is wrong; if the one you can currently see on my route map or the one provided by using the navcam polar projection. That's why I want to wait until we are closer to Epsilon (end of this week?) to gather some precise heading and distance measurements and then refine the route.

I think a simple circle in a vertical plane, intercepting surface in only two points... too simple?

I don't quite get you dilo - it might be an ellipse in general, I'll work it out though.

As you probably are aware from the MOC website, we've done this for both rovers. But it's expensive, in terms of MGS operations, and as Mike pointed out, we do know where the rover is with respect to the MOC images that have been taken.

-Tim.

Here is an example of a truly bizarre vertical stretch of the latest navcam pan, made into a polar. This was mhoward's pan, and I cropped out just a thin strip - maybe 100 pixels or so high - from a 7000 pixel wide pan, made it 3000 by 3000 and did a polar of it. At this level of stretching various distant features become apparent, including crater Delta (as shown on Tesheiner's map), possibly crater Gamma, and a couple of outcrops which are obvious on his route map. Several other dark spots ahead of us are clear as well. This is enough to triangulate a decent position. The flatter an area is, the more useful this sort of trick is.

Also, of course, it shows the distant relief really well. The broad depression north of Victoria, the double horizon west of the current location, etc.

Phil

Click to view attachment

I stand corrected. They'll *never* put you in a box. ...and I learned something valuable about polar projections that should have been, but wasn't previously apparent to me. Thanks.

My first estimate for sol 916 , but i think is correct ...

My route map, updated to sol 916.

Click to view attachment

Everyone got their seat belt strapped on?

Everyone got their image stitchers warmed up?

This is gonna be huge

My latest position.

Phil

Click to view attachment

Here's my estimate of Opportunity's position on Sol 916.

I used the position(s) for Hawking Rock which I'd determined using wide baseline photogrammetry from Oppy's positions on Sols 887 and 904 (see the posting in the Route Map thread). Using AlgorimancerPG I measured the range and azimuth to Hawking Rock from the Sol 916 pancam images, finding an azimuth of 158.3 degrees and range of 95 +/-10 meters. I used this information to predict the Sol 916 position, entirely based upon that azimuth, range, and error, plus also estimated the east-west error in the Hawking Rock position (likely worst case). Here are the results (big file, 4.4 Megs, yes I could make it smaller, but I have other priorities at the moment):

http://www.clarkandersen.com/R1500822_NWVi..._Hawking916.PNG

The colored pixels are the predicted positions as exact offsets from the Hawking predictions. The bounding quadrilateral is an estimate of the error range. The white cross is the current position from Tesheiner's route map, the white circle is an estimate of the position from Phil Stooke's route map (it was harder to match up the exact position on that one).

I'm pretty happy with the results. Tesheiner's position falls neatly within my error region. I may have misplaced Phil's position. This makes me a bit more confident in my Hawking position

Looking forward to the next move.

And here is part of my route map where I activated the layers showing the heading lines I usually use to double-check a sol's position.
The ones for sol 914 were taken from the navcams (using APG) while the lines for 916 are derived from a polar projection of the latest "drive direction" pancam mosaic. There is a clear discrepancy, at least when using this basemap image, and as I said before I'll try to correct this part of the route when we get closer to Epsilon (if we continue driving towards it).

Click to view attachment

At this rate, we'll be at the rim by next weekend

A quick word of appreciation from a long-term lurker, Tesheiner. I'm sure that I'm not the only one who checks this thread religiously every day. I appreciate how much you've refined your MER tracking techniques over the last two years.

The question is what shall we all do once the MER rovers finally grind to a halt? Is there life after MER?

Yep... but it'll have two missing bright spots. I've been, and most everyone else here, has been having a great time watching one of the most interesting adventures in the history of mankind - two, not just one, little machines crawling around on a planet (old moon, it seems) for now more than two years.

It has been a time full of drama, discovery, and surprise. A steady stream of pictures that have amazed and confounded. I'll sure miss it all when these two little machines are finally retired somewhere.

One consolation - there'll be the next set of exploratory vehicles - bigger, faster and packing new abilities that'll keep us glued to the pictures they send to us.

I'm curious as to what software you're using to do the overlays. Photoshop? Or something custom?

One thing that caught me by surprise, and shouldn't have, is that I had expected the hillocks to be useful navigational references once we got out onto the annulus, and of course being about the same height or less as the annulus, they're not. Makes me wish sometimes that the rovers carried sets of little flags that they could plant at various sites, just to provide a subsequent reference target Perhaps a can of fluorescent orange spray paint to mark rocks with on the MSL? (okay, orange might be a bad choice...)

I'm thinking that it's perhaps premature to assume that Victoria is where it ends, at least for Oppy. At this point it is not inconceivable that Oppy may later head off to the east and visit that big crater whose northern rim forms the Twin Peaks. It's not dramatically further away than we've already driven, and Oppy seems pretty healthy. Otherwise there's some interesting geography to the south. And of course there's Spirit and a thorough exploration of the environs of Home Plate coming up. Next of course there's Phoenix, which will almost certainly overlap with MER point (though a relatively brief overlap). MSL seems a long way off though :/

I had no Photoshop available when I started tracking the rovers position last year, so I searched for no $$$ tools and found Serif PhotoPlus. I've been using this suite (I'm currently using v6) since then with the idea to change to Photoshop when possible. Now I have photoshop too, but I found it's not directly possible to migrate the map (not the results I post here, but the original file with *all* the different layers -- background maps, ground features, route, future path, [hidden] polar projections and heading lines, etc.) from one tool to the other one, so I continue using Serif PhotoPlus.



My consolation too, BUT I'm afraid the data availability we have with MER will be unique. I hope they repeat the same successful approach with the next rover but I'm not so sure.

Latest route map (sol 917).

Click to view attachment

From that route map, it looks like Opportunity crossed into some lighter (or differently textured) terrain sometime during sol 917. Is this reflected in the images from that sol?

I thought that was the much smoother terrain that was seen in the rear hazcam images -- flat, no ripples.

Once again, here's my take on the current Oppy position, referenced from my earlier wide baseline estimate of Hawking Rock's position.

[edit: filled-in the rest of the web path]
http://www.clarkandersen.com/R1500822_NWVi..._Hawking917.PNG

Again, the white crosses are from Tesheiner's Sol 916 and Sol 917 route map positions (Tesheiner and I are diverging a bit more than we were yesterday). The colored pixels are offsetted estimates of Hawking's position.

For sol 917, I measure Hawking to be 22.6 +/- 1.8 meters from Oppy, at a bearing of -138.2 degrees azimuth. Unfortunately all I have to work with are navcam images, there are no pancam images of Hawking

This would be a good time for someone to whip-up a planar projection of the Sol 917 navcam panorama. Overlaying that projection onto the route map may finally allow us to clinch Hawking's position to within about a meter, which will have a nice trickle-down effect on prior and subsequent route map positions.

Since there has been some uncertainty about recent positions, I thought I should explain my 916 location. This is how I arrived at it:

Click to view attachment

This is my 916 polar and part of Tesheiner's map. The circled features indicate the rover location - as I see it. These identifications are always a bit suspect. I haven't done 917 yet.

Phil

I notice that JPL's website have updated their maps of Opportunity's progress as far as Sol 914.

======
Stephen

Nice match, Phil.
And when comparing your estimation with the one I've got measuring headings to Epsilon and VC rim features (see red heading lines on this previous post) it looks they fit quite well. That seems to be the right position for sol 916.

For 917, I have a "preliminary" correction based on headings (references are Epsilon and Hawking) and it places the rover about 32m SE of the position plotted on my last route map update.

It looks like we'll reach Epsilon in two sols (919?), then it will be time to "fine tune" the whole route segment since sol 914.

I use Paintshop Pro (PSP), which is full-featured and reasonably-priced. They have a website, and I think that you can download a demo version for evaluation.

--Bill

There's always http://www.gimp.org/

Doug

Wop! good reference. I have a very old version of Gimp.

Rodolfo

a small attempt at checking of telemetry by the way 3D. a method of reprojection of element 3D on the panoramas. If all the parameters of telemetry are good one must have a good match on the panorama.

a small test on “epsilon”
sol 917
site 755J

Click to view attachment
Click to view attachment


I add a news box in the scene near "beacon" …

Click to view attachment

and it goes

Click to view attachment

and finaly a top view

Click to view attachment

the method is not very precice, but I go tried to refine

The lingustic discussion is great and all, but I'm waiting for the next update to the route map! I checked exploratorium and it looks like Oppy drove today, if we're not seeing more hazcams from previous sols. (They do look different than the images from 8/23, at least some of them).

No drive today. The images are the back part of the 917-918 Navcam mosaic.

I think the linguistic discussion, although great and all, would be even more fascinating in a more appropriate thread.

Ah, gotcha. Thanks. I was puzzling over some of the exploratorium images from the forward hazcam. How far are we from the Victoria rim now?

Edit: I'm guessing 220-230 m based on the last route map update (a crude method of using the last 74m drive points like a map scale).

SOL 919 is a driving sol

Just to clarify, 8/24 is not a drive sol? And, we can hope for a new route map on Aug 25?

The fascinating linguistic discussion is continued here;

http://www.unmannedspaceflight.com/index.php?showtopic=3108

It is now the early morning hours of Sol 919. Do you know where your rover is?

After Oppy wakes up, driving will commence, no doubt. That will be sometime on the 25th now, unless I've made a math error.

I moved two new liguistics posts to the linguistics thread, or tried to. Anyway, please take any further linguistics discussion there. Any more linquistics posts here will be summarily moved - and probably lost, due to my lack of expertise in moving posts.

Which raises the question of whether people should be allowed to play with matches if they don't understand how to use them?

======
Stephen

Careful... I have the ability to just 'delete', as well As it happens, it looks like those posts found there way over there safe and sound... just sorted chronologically, I guess.

Before anyone asks... the move today (sol 919) was inconsequential, just a setup for the trenching operation over the weekend. So go enjoy yourselves in the linquistics thread, I guess.

Back on topic.
Tosol (919) move was so short that it doesn't deserve a map update. Pick sol 917 position and add a red dot about 2.5m to the east.

Here's my attempt at the Sols 817-present positions, as well as Bradley aka "Hawking". I'm using something similar to Phil's approach here, building from my prior wide baseline position for Hawking. I used the current version of AlgorimancerPG to create a 100 meter square planar projection of the navcam images, reduced that projected image to the same 2 pixels per meter as the route map image, and used Corel Photopaint to overlay the projected images onto the map image, varying the transparency as I shifted it back and forth until achieving a good match with the background details. This was quite challenging, as at this scale it is difficult to find corresponding details to match-up. Ultimately I believe that I have locked the position to within 1 meter of the correct location based upon multiple matching foreground and background features which lost alignment upon a 1 pixel shift in any direction. This may be wishful thinking on my part (the position for Bradley is very close to the yellow cross position found vial wide baseline PG). On the same map view I also attempted to do the same for Sol's 816, 814, and 812. These positions were much tougher to match up, Sol 814 in particular is mostly a guess (some bad orientation information for some images in both Sol 814 and 816, I think). Anyway, at best these positions may be accurate to several meters (for Sol 816), perhaps double that error for Sol 812, and Sol 814 may be off by tens of meters for all I know. I haven't made any effort to correlate these positions with Phil Stooke's or Tesheiner's positions. At this point I'm leaning towards backing-off and leaving the mapping to the experts, I just particularly wanted to get a good fix on Bradley's position.

Click to view attachment

Oh yes, for anyone else who might like to play "Find the Rover", here are the 2-pixels-per-meter plane projection images (same scales as the route maps) I used to do the same:

912
Click to view attachment

914
Click to view attachment

916
Click to view attachment

917-918
Click to view attachment

919-Present
Click to view attachment

Don't cut yourself short. It looks like your doing an excellent job! This looks like something that Tesheiner and Phil had been doing when we were traversing the drifts around Erubus. But I think that Tesheiner was eye-balling it with features on the the ground, whereas you are creating an overlay to do the match. Clever!
Keep it up! The more the merrier I say

Nice work, algorimancer. This route map is quite familiar and strongly recall also "official" JPL maps...
Anyway, in order to have a better surface visibility in each vertical projection, I suggest you to "equalize" illumination in the various directions. Moreover, Sol914 projection show the horizon and probably needs a tilt correction (this is partiall true also for Sol916).

Thank you. Equaling illumination is a good idea, I suppose I was being lazy. The problem with the horizon on Sol's 914 and 916 is apparently related to the reset that occured around that time, throwing-off the rover's orientation quaternion. I'm not sure that there's anything I can do about that, as my technique is entirely dependent upon that information (I have no means of adjusting per the horizon, I'm completely dependent upon the camera pointing information and rover orientation quaterion, unlike your POV-Ray approach). It appears to only affect the images from the northeast quadrant, which I don't understand. Anyway, I would be inclined to trust Phil Stooke's and Tesheiner's pre-Sol 917 positions more than my own. You might try a planar projection using POV-Ray for the Sol 919 position, and see whether you come up with an equivalent position.

[edit]Upon further investigation I'm beginning to suspect that the projection went awry on (at least) the Sol 914 and 916 sites, perhaps the others as well. Looks like it picked-up some tilt from somewhere (in my code I think). Doesn't seem to invalidate the Sol 917+ positions, but the projections are definitely problematic.

[edit again]Nope, my code seems fine. Looks like the problem is either that the rover is sitting on a 1 1/2 degree slope on those two occasions, or the rover orientation quaternion is off by that much on those occasions. Conceivably I could add a correction based on measured horizon elevations at particular azimuths.

An updated route map for sol 929.

Click to view attachment

Note that I corrected the route path since (and including) sol 914 drive. Those points were all strictly based on the own rover's movement data and were day by day drifting from what we could calculate by triangulation.
Now that Oppy is right beside Epsilon Emma Dean, it was quite easy to "stretch" the route to fit with the real position as of sol 929. I double-checked the "new" positions for those earlier driving days with the polar projections for the respective sols and they match quite well.

Now I think it's time to shift the map.

guess there's a little typo on your map: "920" should be "930".

Oooops, thanks for the info general.
I've just updated my previous post with the corrected driving sol, which btw was 929 and neither 920 or 930.