Printable Version of Topic

Posted by: elakdawalla Mar 7 2006, 06:58 PM

I am building some web pages about MRO (we're hosting the team website for Mars Climate Sounder) and am looking for a tool that I can use to draw some pictures of the shapes of orbits and ground tracks. I know that there are fancy tools out there, like SOAP, but I am hoping that there are some tools out there that are a little easier to use that I can use to generate some simple diagrams. For instance, I'd like to create one diagram on a flat map projection (simple, mercator, whatever) that shows the 17-day ground track repeat for MRO, and one with Mars as a sphere showing the position of the near-polar orbit, maybe one comparing the orbits of MRO, MGS, and Odyssey, even maybe generate a bunch of diagrams to make an animation showing how as MRO goes around Mars its orbit is sun-synchronous. I know I could do all of this with SOAP but I was never able to use SOAP without a lot of hand-holding from a very patient person at JPL. I would appreciate any advice or help on simpler tools to visualize orbits at Mars!

--Emily

Posted by: AlexBlackwell Mar 7 2006, 07:18 PM

I'm not sure if it helps (or maybe I don't understand precisely what you are looking for) but perhaps Noel Gorelick's http://jmars.asu.edu/, which has a groundtrack layer, might be useful.

Posted by: mcaplinger Mar 7 2006, 07:32 PM

It's somewhat surprising, but I don't know of a tool short of SOAP or STK, both of which have pretty steep learning curves, that do both of the things you want. Most targeting tools only draw groundtracks on maps, not on pretty 3D spheres. For public talks I've used Starry Night and faked the orbital elements. I've heard good things about Celestia, but I don't think it can use SPICE files directly, which you would really want for the groundtrack walk and the sun-sync orbit (both of which would require J2 perturbations to model correctly.)

Posted by: AlexBlackwell Mar 7 2006, 08:02 PM

I haven't looked at http://www.stk.com/products/desktopApp/stkFamily/ in quite a while. I wasn't aware that STK/Advanced VO could run on Windows-based platforms.

Posted by: mcaplinger Mar 7 2006, 08:28 PM

I didn't know it ever ran on anything but Windows.

STK probably isn't a great choice. I forget if the free version supports central bodies other than Earth. The SPICE file support, at least when I looked at this a few years ago, was fairly rudimentary and came in an expensive add-on package, like most other STK features.

Posted by: AlexBlackwell Mar 7 2006, 08:32 PM

Long ago, I was under the impression that it ran only on UNIX-based platforms, though I'm probably confusing it with another app run exclusively on Sun Micro workstations. At any rate, when I checked out the AGI website, I saw that STK could run on both Windows and UNIX.

Posted by: mcaplinger Mar 8 2006, 12:54 AM

Absolutely right, I stand corrected. At any rate, STK isn't likely to do what you want. It's too bad that no one has written a simple, general tool to do this stuff. Unfortunately, PDS doesn't seem to have released the source for their planet viewer tools -- http://pds-rings.seti.org/tools/index.html -- because that would be close to what one might want.

Posted by: elakdawalla Mar 8 2006, 12:57 AM

Thanks for the comments and suggestions. Looks like I may have to try pounding my forehead against SOAP to get what I want.

--Emily

Posted by: AlexBlackwell Mar 8 2006, 01:24 AM

I'm also surprised there isn't a publicly available tool to do this. I do know that Cassini Mission Planning (MP) uses a software package named Science Opportunity Analyzer (SOA), which, I believe, could supply a lot of the features Emily is looking for. Unfortunately, even the SOA User's Guide is ITAR-protected (i.e., as you know, under the category of I-could-tell-you-but-then-I'd-have-to-kill-you), so I'm sure the SOA source code isn't in the public domain, either. And like STK, SOA runs under both Windows and Solaris.

Posted by: mcaplinger Mar 8 2006, 05:24 AM

I looked around but it looks like Aerospace Corp has continued to make it difficult to get SOAP, and I don't have a copy here, or I would try to help you out.

For the groundtrack walk, I could crank something with our targeting software.

I installed Celestia and it's not too bad, but for the fact that it only processes Keplerian orbital elements, which for spacecraft will stay accurate for a few days at best. And translating Mars orbits to the right coordinate system will be a somewhat painful bit of tedium (if you recognize the phrase "IAU vector and equinox of epoch", you know what I mean).

Here's an example of what Celestia shows for MGS and Odyssey. It's pretty, but the positions (based on some config files I found on the net) are fiction. And Celestia appears to save JPEGs flipped left-for-right. Oh well, so close!

Posted by: helvick Mar 8 2006, 08:37 AM

Celestia shouldn't be that far off on the instantaneous locations of the spacecraft\solar bodies provided you are using accurate spacecraft xyz files but the orbital tracks have been an issue IIRC. There was a problem at one time with the texture maps for Mars where the default Mars surface texture was 180deg out of line with reality - that was in a fairly early version so it shouldn't be a problem with anything recent. I've never seen the inverted jpeg problem before - spooky.

Posted by: jmknapp Mar 8 2006, 12:23 PM

I've also been thinking about the best way to visualize MRO's operations--not easy. As for the 17-day groundtrack, here's one approach (click for larger image and to see the faint ground tracks):

http://marsrecon.com/mrotracks.jpg

If you zoom in you can see that MRO (the red marker) is just beginning to overlap just to the left of the tail of the track, where it started 17 days earlier. It's interesting to note that the field of view of HiRISE at this scale is only 1/2th the width of the groundtrack line. So even the CTX FOV is just a few line widths (pixels). This was generated from a SPICE kernel & so should be fairly accurate. Did you want to see that on a sphere? Still, it's kind of busy.

It might be neat to see a view towards the "front" of MRO, along the groundtrack--to see where it's going so to speak. Something like:



I made a (warning: large file) animation along those lines, showing MRO traversing from the south pole to the north:

http://cboh-t.cboh.org/~jmk/hirise.mpg

Most ideas I have get quickly out of hand in terms of file sizes, bandwidth and processor demands, etc.!

Posted by: mcaplinger Mar 8 2006, 02:39 PM

Certainly, but because there's no J2 perturbation modeling you would have to update the elements every 2-3 days to maintain their accuracy, which is a bit of a hassle and makes making an animation over the span of months a real pain.



Most of the plots of the repeat cycle I've seen zoom in and show orbit 1, orbit 2, and orbit N, where N in the length of the repeat (327 orbits nominal for MGS, IIRC, or a little under a month).

Of course, it's important to note that the repeat cycle is just an idealized case which usually doesn't happen exactly that way in practice. We never do MOC targeting a month out because the predicted ephemeris is just not accurate enough to support that.

Posted by: mcaplinger Mar 8 2006, 05:26 PM

No, wait a minute, you're right. I was assuming that the only way to put the orbit into Celestia was to supply orbital elements. I now see that there's an option to use "SampledOrbits" which are just the XYZ locations at some time spacing. Assuming that Celestia wouldn't choke with, say, a few months' worth of 5-minute or 1-minute spacing, one could write a program to extract that from the SPICE files. Then there's just the wrinkle of figuring out the coodinate system.

Posted by: jmknapp Mar 8 2006, 05:41 PM

I guess that over a few months one would animate the shape/plane of the orbit rather than MRO itself which would be going around more than 12x per day. So what would such an animation look like? If the observation point was fixed above the nominal value of local mean time, probably the MRO orbit plane would oscillate back and forth over the year as Mars goes through its equation of time and true local time varies.

Posted by: elakdawalla Mar 8 2006, 05:48 PM

This map is great -- it is busy but it should be, it illustrates the network of the ground tracks nicely. Thanks also for the details about the relative scales of the FOVs.

Your video is really cool -- too big for my website but really cool!

--Emily

Posted by: helvick Mar 8 2006, 06:05 PM

Sorry I wasn't clear about that - the Celestia forum guys have some tools for this.
http://shatters.net/forum/viewtopic.php?t=6188&highlight=&sid=ea8de8fbcc7c6c9c5d0da0a3fd135d70

Posted by: jmknapp Mar 8 2006, 07:00 PM

Emily -- glad you can use the map.

Helvick -- do the Celestia xyz files use J2000 coordinates?

Posted by: mcaplinger Mar 8 2006, 07:19 PM

I couldn't find any documentation on this. The orbital elements are pretty clearly given in some central-body-equatorial system, but where the X vector is pointed wasn't clear to me. While possible, it seems unlikely that the Celestia developers settled on the IAU_MARS frame (IAU vector and Mars equinox of 2000)

BTW, there's a small MGS Mars map at http://www.msss.com/allmars.jpg (1440x720) so you don't have to keep using that old Viking map I'll see if I can't put a bigger one out soon.

Posted by: helvick Mar 8 2006, 07:26 PM

Yes - The data (with some reformatting) is the same as that produced by http://ssd.jpl.nasa.gov/horizons.cgi although on the only occassion that I used it I used the email option as per the following:

Here's a sample template for an e'mail message to request xyz coordinates
from Horizons for use with Celestia. You will have to edit the returned
message into the format that Celestia uses.

The message must be mailed to the address "horizons@ssd.jpl.nasa.gov"
The subject must be "JOB" ( do not include the quotes in either )

The body of the message should consist of the following text which
is between the lines of =====. Don't include those lines.
The lines starting with "!$$" are required, however.
Replace the fields ending in "here" with appropriate values.
Don't touch any of the other lines until you've read and understood
Horizons' documentation.
==========================================================
!$$SOF (ssd) JPL/Horizons Execution Control VARLIST
EMAIL_ADDR = 'your address here' ! 'you@there.com'
COMMAND = 'object name or id# here' ! e.g. 'TRITON'
OBJ_DATA = 'NO'
MAKE_EPHEM = 'YES'
TABLE_TYPE = 'VECTORS'
CENTER = '@10'
REF_PLANE = 'ECLIPTIC'
SITE_COORD = '0,0,0'
START_TIME = 'starting date here' ! e.g. '1989-Aug-23 09:00'
STOP_TIME = 'ending date here' ! e.g. '1989-Aug-27 09:00'
STEP_SIZE = 'sample period here' ! e.g. '10m' or '2d'
REF_SYSTEM = 'J2000'
OUT_UNITS = 'KM-D'
VECT_TABLE = '1'
VECT_CORR = 'NONE'
TIME_DIGITS = 'MIN'
CSV_FORMAT = 'NO'
VEC_LABELS = 'NO'
ELM_LABELS = 'NO'
R_T_S_ONLY = 'NO'
CA_TABLE_TYPE= 'EXTENDED'
CALIM_SB= '0.1'
CALIM_PL= '.1, .1, .1, .1, 1.0, 1.0, 1.0, 1.0, .1, .003'
!$$EOF++++++++++++++++++++++++++++++++++++++++++++++++++++++
==========================================================

Posted by: jmknapp Mar 8 2006, 10:17 PM

You mean for the map that has the ground tracks? I used a Celestia map done by praesepe the original of which looks like the one you mention--must be the one it's based on. For the animation I used your 256 pixel/deg maps.

Posted by: mcaplinger Mar 9 2006, 12:39 AM

Note that Horizons is outputting everything in heliocentric XYZ. If you did that for a Mars-orbiting spacecraft, then if Celestia's idea of where Mars is would have to match JPL's exactly, or the orbit would come "unstuck" from the planet. I'm pretty sure that the XYZ coordinates that Celestia uses should be Mars-centered. I'll have to get the source to see if it's oriented in J2000, but I seriously doubt that it is.

Posted by: Sym05 Mar 9 2006, 01:19 PM

Try to investigate jpleph.dat for Celestia, I found it in a thread about New Horizons to Pluto add-on. They had a similar problem.
http://www.shatters.net/forum/viewtopic.php?t=8730


In this topic the author of Celestia, Chris Laurel, give some details: http://www.shatters.net/forum/viewtopic.php?t=4273&postdays=0&postorder=asc&start=15

Have a look at http://www.shatters.net/~claurel/celestia/files/jpleph/

There are three JPL ephemeris files. The .405 files cover a timespace of 50 years. The .406 file covers 300 years from 1800-2100. The max interpolation error of the DE406 ephemeris is slightly greater (up to 25m versus 1mm for DE405), but this shouldn't matter for Celestia. To experiment with the JPL ephemerides, just copy one of the files (probably unxp1800.406) into Celestia's data directory as jpleph.dat. Then, change the CustomOrbits in solarsys.dat from vsop87-<planet> to <planet>-jpl. There's still a slight discrepancy between the positions Horizons returns and the positions I compute from the JPL ephemerides--for Jupiter, it's up to 200km. At this point, I suspect the problem may be related to a slightly different time standard.

Posted by: mcaplinger Mar 11 2006, 03:54 AM

I found the following about Celestia's SampledOrbits:

"The origin is the center of the parent object, whether it's a star or a planet. If the parent is a planet, the +z axis is the planet's north pole and the xy plane is the planet's equatorial plane at the reference epoch J2000. The x axis points at the ascending node of the equatorial plane on the ecliptic."

I believe this is the same inertial frame as the NAIF Toolkit's MARSIAU frame (not to be confused with the rotating IAU_MARS frame, alas.)

edit: for those of you still reading this obscure discussion, I think I was wrong: the Celestia frame is what we usually call Mars Mean Equator and Equinox of J2000, which is different from MARSIAU. Back to the drawing board.

Posted by: Mike Loucks Dec 7 2006, 10:04 PM

I'm sorry I didn't see this until now. STK easily does all of what you want.

The statement that "STK isn't likely to do what you want" is way off. STK does all of this stuff, and in my opinion does it better than any of the other tools mentioned here.

It can absolutely use SPICE files, although it doesn't have to because it has it's own numerical integrators that will create a real ephemeris if you like.

The Astrogator module of STK has flown libration point trajectories at GSFC and comes from the GSFC 'Swingby' code that was used to fly Clementine, WIND, SOHO and Lunar Prospector.

STK does ground tracks in 2D and 3D orbits in any frame you can think of. We regularly use it to show planetary fixed, and inertial frames, along with spacecraft relative (formation flying) and rotating libration point frames.

Repeating ground track stuff is easy, at any planet. Animations are also easy. STK runs on Unix and PC platforms.

That said, STK isn't free (like Celestia). But if you work for an organization that owns it, you can do an awful lot of stuff.

Posted by: mcaplinger Dec 7 2006, 10:21 PM

By that I meant the free version of STK. I don't think many people reading this will be able to afford the needed add-on modules to read SPICE files; last I looked they were well over $10K.

BTW, the current version of Celestia can read SPICE SPK files. The support's not all there yet, but it's improving rapidly.

Posted by: Hari Mar 13 2007, 02:48 PM

You can also try http://xplanet.sourceforge.net. It can draw satellite orbits, ground tracks in many different projections, and read spice kernels.

See http://xplanet.sourceforge.net/Gallery/20040222_mars/ page for an example.