Mars Rover

[AZRedhawk44]

http://www.youtube.com/watch?v=FZYnIsLNz3c&feature=endscreen

Anyone know what the theory is behind the rover's pathfinding and rangefinding and mapping capabilities?

It can't be GPS driven since there isn't a GPS network of satellites around Mars like Earth.

Is it waypoint driven by humans at a mission control center, in smallish increments like every hundred yards or less?  Does it have rangefinding capabilities and obstacle or pit detection, or obstacle avoidance logic?

[geronimotwo]

maybe it just keeps its sensors peeled for the shiny yellow stuff.     does mars have magnetic pole?

[birdman]

It navigates via range finding, image matching, sun tracking, and terrain mapping.  Since it is visible by the orbiter, the resulting position can be refined over time as well.

[AZRedhawk44]

Also, I gotta say I am SO impressed by the lander.  Robotic rocket-hover payload deploying device?  Wow!

That thing would look at home in a 23rd century "Alien" movie.  The notion that someone built it to auto-deploy a small car (AFTER being jettisoned from a falling parachute at hundreds of miles an hour!) on the surface of Mars from 300 million miles away is just damned impressive.

The only shame of it is that the device was intentionally crash landed far away from the rover itself, rather than gently landed a few hundred yards or a couple miles away.  Imagine if it could be refueled and used to pick-up the rover again, to help it move across a large canyon or other massive obstacle, or to summit an otherwise un-climbable mesa.

[kgbsquirrel]

Does the rover include a star-tracker?

[birdman]

Does the rover include a star-tracker?

I don't believe it has a built-for-the purpose one, but a star tracker is camera and software.  A star tracker doesn't do much good, it just tells you attitude, something you can get easily with internal accelerometers.  True, you could estimate longitude and latitude with it, but the signal from the orbiter and the sun provide sufficient references. 

From what I understand, local nav is done with imaging and ranging, large scale nav by tracking it with the orbiter.

[Ben]

From what I understand, local nav is done with imaging and ranging, large scale nav by tracking it with the orbiter.

In the late 90's we were using COBRA-TAC dive trackers for nearshore bathymetry mapping of smaller areas like shipwreck sites. Once we programmed in a geodetic starting point, which we got via differential GPS, the tracker used range and direction finding to convert everything to lat/long for diver navigation and waypoint marking, while at the same time using active acoustic signal imaging for on the fly high rez 3D seafloor mapping. I would imagine the rover is using a similar (though well-advanced) method for its local navigation and terrain memory.

[kgbsquirrel]

A star tracker doesn't do much good, it just tells you attitude, something you can get easily with internal accelerometers.

Using two or more star trackers can actually give you accurate positioning data instead of just attitude. As an example the LGM-118 peacekeeper ICBM used them for passive in-flight position and navigation updating (part of the reason it was able to achieve 300 foot CEP after traveling 9000 nautical miles down range.)

But I suppose for ground use you could just get by with a really accurate chronometer and some sort of electro-optical sextant.

[Phyphor]

Curiosity is controlled from earth in a sense.  Back at earth, the Rover Planners examine photos taken from Odyssey (the orbiting mars mapping probe / signal relay satellite,) and from the rover itself.  A 3D "mesh" (kind of like a level in a 3D game) is generated, and the basic course is plotted and "sequenced."  Once a sequence has been generated, it's rechecked, compiled, and transmitted to the rover.  Sequences cover everything from driving, photographing, use of the AXPS and other scientific equipment.

The rover does have some autonomy.  If they tell it to basically "go over here," it'll do it's best, but will use several instruments to determing tilt, slippage, and any other hazardous conditions it may run into.  If it finds something it doesn't know how to get around, it'll stop and contact Odyssey (which will then relay the rover's current status back to Earth, hopefully so the RP's can figure out how the hell to get it out of whatever fix it's in.)

The same basic software / strategies were used to great success for both Spirit and Opportunity, which both lasted far beyond their 90 sol assumed design limits.
(Opportunity is actually still functioning, and is currently studying what looks to be evidence of ancient seas on the other side of the planet, )

[birdman]

Using two or more star trackers can actually give you accurate positioning data instead of just attitude. As an example the LGM-118 peacekeeper ICBM used them for passive in-flight position and navigation updating (part of the reason it was able to achieve 300 foot CEP after traveling 9000 nautical miles down range.)

But I suppose for ground use you could just get by with a really accurate chronometer and some sort of electro-optical sextant.

Not quite.  The star trackers allowed for better attitude reference to better utilize the accelerometers on the space stable platform.  So position accuracy was indirect.

[kgbsquirrel]

Not quite.  The star trackers allowed for better attitude reference to better utilize the accelerometers on the space stable platform.  So position accuracy was indirect.

Hmm. You may be right but I remember it differently. In any case I don't have access to intelink anymore so I can't go back and recheck my source. 

[RoadKingLarry]

Also, I gotta say I am SO impressed by the lander.  Robotic rocket-hover payload deploying device?  Wow!

That thing would look at home in a 23rd century "Alien" movie.  The notion that someone built it to auto-deploy a small car (AFTER being jettisoned from a falling parachute at hundreds of miles an hour!) on the surface of Mars from 300 million miles away is just damned impressive.

The only shame of it is that the device was intentionally crash landed far away from the rover itself, rather than gently landed a few hundred yards or a couple miles away.  Imagine if it could be refueled and used to pick-up the rover again, to help it move across a large canyon or other massive obstacle, or to summit an otherwise un-climbable mesa.

I'm still kind of geeked out about the whole operation. Restores some small degree of hope for humanity.

[birdman]

Hmm. You may be right but I remember it differently. In any case I don't have access to intelink anymore so I can't go back and recheck my source. 

I based my logic on designing spacecraft, while a true stellar sextant (multiple or wide FOV star trackers integrated over time) can provide some positional accuracy (basically, solving orbital or sub-orbital equations of motion), it take integration time--you need to move over a substantial angle of the orbit (angle much larger than the resolution of the tracker) so it doesn't help when you need the accuracy to operate a PBV, between drop-offs.   But it DOES help substantially with attitude reference of a space-stable platform when there isn't an apparent gravity vector (ie free-fall) or when the attitude of the launch platform may have errors (ie a non-fixed launch, like a sub).  I believe the latter is why a trident could be aided by stellar-aided inertial...but I'd have to check if it actually does,

[RoadKingLarry]

when the attitude of the launch platform may have errors (ie a non-fixed launch, like a sub).

Submarine navigation was was pretty precise 20 years ago, pretty much down to a 3 meter or so accuracy. I can only guess that it is at least as accurate today. Plenty precise enough for a nuke.

[kgbsquirrel]

Submarine navigation was was pretty precise 20 years ago, pretty much down to a 3 meter or so accuracy. I can only guess that it is at least as accurate today. Plenty precise enough for a nuke.

I wonder if that could have been a good re-use of the decommed peacekeeper parts. They had an ultra-accurate inertial navigation system.

http://nuclearweaponarchive.org/Usa/Weapons/Airs.html

Perhaps it could  be installed in the subs instead?

[birdman]

I wonder if that could have been a good re-use of the decommed peacekeeper parts. They had an ultra-accurate inertial navigation system.

http://nuclearweaponarchive.org/Usa/Weapons/Airs.html

Perhaps it could  be installed in the subs instead?

Sub ones are better.

The issue isn't the quality, it's the integration time.  Angular errors build up over time. 

[ArfinGreebly]

Submarine navigation was was pretty precise 20 years ago, pretty much down to a 3 meter or so accuracy. I can only guess that it is at least as accurate today. Plenty precise enough for a nuke.

Heck, I'd take 3 meters, even for a MOAB.  I'd even settle for 50 meters.

After all, it's horseshoes, right?

[kgbsquirrel]

Sub ones are better.

 

*scoots chair closer* Do tell. I was under the impression that the AIRS was pretty much state of the art for this sort of thing.

[birdman]

*scoots chair closer* Do tell. I was under the impression that the AIRS was pretty much state of the art for this sort of thing.

AIRS is pretty good.  But you can do better when you are less space and weight constrained.

Btw, side note, the best gyros in the world are the gravity probe B gyros.

[280plus]

That is all very nice but can it make a decent cup of coffee?

[Phyphor]

It can heat it up, using either the laser or just boiling the water off the RTG pack....

[AJ Dual]

AIRS is pretty good.  But you can do better when you are less space and weight constrained.

Btw, side note, the best gyros in the world are the gravity probe B gyros.

IIRC,  quartz spheres so perfect if they were  expanded to the size Earth, the biggest bump would be something like .25 inches high or whatever...

Seven orders of magnitude less drift than the best .mil gyros.

[birdman]

IIRC,  quartz spheres so perfect if they were  expanded to the size Earth, the biggest bump would be something like .25 inches high or whatever...

Seven orders of magnitude less drift than the best .mil gyros.

Yup, and superconducting (niobium coated), spun up with gas jets.

[RoadKingLarry]

Electrostatically supported Beryllium Sphere spinning really fast.


http://www.militaryaerospace.com/articles/2006/12/navy-opts-to-continue-supporting-precise-navigation-system-for-ballistic-missile-submarines.html

http://ieeexplore.ieee.org/xpl/login.jsp?tp=&arnumber=670128&url=http%3A%2F%2Fieeexplore.ieee.org%2Fiel4%2F5490%2F14769%2F00670128.pdf%3Farnumber%3D670128

 When I was doing it the ESGN systems were pretty freaking accurate. They were also updated as needed by external position fixes from NAVSAT, OMEGA , visualand towards the end of my time in, GPS as well as calibrated to the known pier side position.

[Scout26]

IIRC, Mars doesn't have magnetic poles, one of the reasons why it's atmosphere is so thin.






Yep,

http://www-ssc.igpp.ucla.edu/personnel/russell/papers/mars_mag/

Magnetic poles protect the atmosphere from solar winds.