Red Dragon Discussion Thread (1)

[baldusi]

They may be looking for an exceptionally flat spot with ice under it. Less flexibility is the price you pay for not going for something like MSL's system.

But what's the CEV of this? 20km on the short side of the ellipsis? Remember, no GPS here. You can either do inertial or terrain recognition. Plus whatever the star tracker give you as initial position. I don't know much, but it sounds to me that a skip reentry isn't the most precise maneuver. After that, you have to actually look down to land.
I work close to glaciers. Ice is a very dangerous, full of cracks. And it's very traitorous since it can break easily without waring. Even HiRISE lacks the resolution to analyze those dangers. And the stock Dragon is not designed for that kind of landing. It's big, heavy and very tall. It will need to land on "nice" terrain.

[apace]

Even the robotic Dragon with full return and pin point landing capability, will make use of GPS and will land on a cement landing pad. Here we are talking about flying by inertia, recognizing the terrain, and deciding, with no external aid whatsoever (save the MRO maps), where there's a patch of terrain clean and leveled enough to actually land.
We are talking about 3D terrain recognition on real time. That's stuff that even the Google car is not ready for. There's a reason that Curiosity uses those huge wheels and suspension, and the Sky Crane works the way it does. Its to land on very unfriendly terrain.

I can remember that I have seen such a research project recently on the Nasa video archive. And from my military experience, such system exists, especially if you have no other driving vehicles around, which you also need to add in your calculations.

[Robotbeat]

Skip entry is guided entry, though. And that allows the possibility of more precision than unguided entry. Remember how Dragon splashed down within about a mile of its intended spot.

Viking 1&2, Pathfinder, Spirit, Opportunity, and Phoenix were all unguided (though some had a lifting entry). MSL is the first one we've sent with guidance. Once we have more precise monitoring of the Martian atmosphere (MAVEN, launch set for 2013), better-than-MSL entry guidance should be possible. Also, remember that Red Dragon would do EDL without parachutes, so low-level winds are much less of a factor.

[baldusi]

Even the robotic Dragon with full return and pin point landing capability, will make use of GPS and will land on a cement landing pad. Here we are talking about flying by inertia, recognizing the terrain, and deciding, with no external aid whatsoever (save the MRO maps), where there's a patch of terrain clean and leveled enough to actually land.
We are talking about 3D terrain recognition on real time. That's stuff that even the Google car is not ready for. There's a reason that Curiosity uses those huge wheels and suspension, and the Sky Crane works the way it does. Its to land on very unfriendly terrain.

I can remember that I have seen such a research project recently on the Nasa video archive. And from my military experience, such system exists, especially if you have no other driving vehicles around, which you also need to add in your calculations.

I never said it can't be done. I would also think about doing a small sat that can point a laser to the desired point of landing, for example. Laser guided Dragon, I like the ring of it.
But neither of this things is stock Dragon, nor easy, nor cheap. They stated that they wanted to do this for 150M LV+capsule! The sheer development that's needed would be huge. And I doubt the military supplier will license their technology for cheap. Now, if you get DoD to tell them to do it to perfect it, that's a whole different ball game. But that's not SpaceX usual MO.

[apace]

I never said it can't be done. I would also think about doing a small sat that can point a laser to the desired point of landing, for example. Laser guided Dragon, I like the ring of it.
But neither of this things is stock Dragon, nor easy, nor cheap. They stated that they wanted to do this for 150M LV+capsule! The sheer development that's needed would be huge. And I doubt the military supplier will license their technology for cheap. Now, if you get DoD to tell them to do it to perfect it, that's a whole different ball game. But that's not SpaceX usual MO.

I agree with you, that such a mission including R&D costs much more than 150M. But I think we should take the 150M more like the figure, how much a plain Dragon and Falcon H costs and all additional features are on top. Paid by Nasa of course ;-)

[manboy]

I have a few questions:
Are Dragon's legs rated for unleveled terrain?

As the paper plays with the numbers and features of the enhanced crew version of Dragon (with propulsive landing), your questions can only be answered from one inside Dragon development. But about autonomous landing capability, there should be enough systems around now which can be used for this.

Even the robotic Dragon with full return and pin point landing capability, will make use of GPS and will land on a cement landing pad. Here we are talking about flying by inertia, recognizing the terrain, and deciding, with no external aid whatsoever (save the MRO maps), where there's a patch of terrain clean and leveled enough to actually land.
We are talking about 3D terrain recognition on real time. That's stuff that even the Google car is not ready for. There's a reason that Curiosity uses those huge wheels and suspension, and the Sky Crane works the way it does. Its to land on very unfriendly terrain.

ALHAT is already being developed, maybe it could be used. Although I'm not sure how you would integrate it into Dragon and what the estimated system weight is.

EDIT: Autonomous Landing and Hazard Avoidance Technology (ALHAT)
Here's the project website http://alhat.jpl.nasa.gov/
And a Mars EDL presentation that mentions ALHAT.

[jongoff]

I have a few questions:
Are Dragon's legs rated for unleveled terrain?

As the paper plays with the numbers and features of the enhanced crew version of Dragon (with propulsive landing), your questions can only be answered from one inside Dragon development. But about autonomous landing capability, there should be enough systems around now which can be used for this.

Even the robotic Dragon with full return and pin point landing capability, will make use of GPS and will land on a cement landing pad. Here we are talking about flying by inertia, recognizing the terrain, and deciding, with no external aid whatsoever (save the MRO maps), where there's a patch of terrain clean and leveled enough to actually land.
We are talking about 3D terrain recognition on real time. That's stuff that even the Google car is not ready for. There's a reason that Curiosity uses those huge wheels and suspension, and the Sky Crane works the way it does. Its to land on very unfriendly terrain.

ALHAT is already being developed, maybe it could be used. Although I'm not sure how you would integrate it into Dragon and what the estimated system weight is.

EDIT: Autonomous Landing and Hazard Avoidance Technology (ALHAT)
Here's the project website http://alhat.jpl.nasa.gov/
And a Mars EDL presentation that mentions ALHAT.

I should also mention that Boeing had a related scanning-LIDAR-based project for Precision Landing and Hazard Avoidance that they had a contract for under the H&RT funding that Griffin canceled when he got into office to fund Ares-I...

~Jon

[jongoff]

I found it. Here is the presentation.

Thanks! What's very interesting is this:

Fire launch abort motors supersonically and use them for remainder
of descent.

They are stating that they can fire their SuperDracos while in supersonic flow? Wasn't that completely out of question?

Nope.  IIRC, NASA had initially ruled supersonic retrobraking out back during Viking days because they had tried a centrally-located engine, and wind-tunnel tests of that had shown some bad coupling with the flow-field.  Doing a central engine for retrobraking was both very unstable and the rocket plume actually reduced the aerodynamic drag, increasing the amount of propellant needed. 

Newer work (in the past 5 years or so) is hinting that mounting multiple engines around the periphery could not only stabilize things, but also augment the aerodynamic drag instead of diminishing it.  It was in one of Bobby Braun's last papers before he took the Chief Technologist job.  Wouldn't be surprised if that research was part of why SpaceX chose the configuration they did for Dragon.  Cause for a purely launch abort and vertical landing situation, there are some benefits to having a single, centrally located engine.

So, while supersonic retrobraking hasn't been flight proven yet, and is thus still low-TRL, it sounds like the initial reasons for concern may now be resolved, so it's well worth trying to investigate the idea further. 

~Jon

[jongoff]

Looks like Dragon can land a ton on Mars

Pigs can also fly.

Of course they can -- we call them "pig-eons" ...

Oh, pigs can definitely "fly", given a sufficient sized trebuchet, cannon, or rocket to strap them onto...

~Jon

[2552]

Newer work (in the past 5 years or so) is hinting that mounting multiple engines around the periphery could not only stabilize things, but also augment the aerodynamic drag instead of diminishing it.  It was in one of Bobby Braun's last papers before he took the Chief Technologist job.  Wouldn't be surprised if that research was part of why SpaceX chose the configuration they did for Dragon.

Would that be this paper?

http://www.ssdl.gatech.edu/papers/conferencePapers/IEEE-2008-1246.pdf

[jongoff]

Newer work (in the past 5 years or so) is hinting that mounting multiple engines around the periphery could not only stabilize things, but also augment the aerodynamic drag instead of diminishing it.  It was in one of Bobby Braun's last papers before he took the Chief Technologist job.  Wouldn't be surprised if that research was part of why SpaceX chose the configuration they did for Dragon.

Would that be this paper?

http://www.ssdl.gatech.edu/papers/conferencePapers/IEEE-2008-1246.pdf

It might be.  Timing sounds about right.

~Jon

[Proponent]

Is ignition of a retro engine in supersonic flow likely to be significantly more difficult than operating such an engine in supersonic flow?

[Robotbeat]

Is ignition of a retro engine in supersonic flow likely to be significantly more difficult than operating such an engine in supersonic flow?

For a hypergolic engine, it's probably not too much more difficult (they tend not to have problems getting started...). For non-hypergolic, the initial conditions will be different, so you'll have to design your igniter (etc) with that in mind.

[simonbp]

They may be looking for an exceptionally flat spot with ice under it.

Which is basically the same criteria they had for Phoenix. Meaning, the landing site would likely be one of the runner-up sites for Phoenix. The difference is, Phoenix didn't have HiRiSE 1-m scale images until right before the selection was made, whereas Red Dragon would have those images right from the start.

Most of Mars is pretty flat; Viking 2 and Pathfinder both happened to find some of the most boulder-rich areas of the planet (because both targeted apparently alluvial structures).

EDIT: Here's a panorama of the Phoenix site for reference; it's plenty flat enough for a Dragon to land on.

http://rhea.la.asu.edu/image_gallery/image.php?ipf=3207

[NotGncDude]

I have a few questions:
Are Dragon's legs rated for unleveled terrain?

As the paper plays with the numbers and features of the enhanced crew version of Dragon (with propulsive landing), your questions can only be answered from one inside Dragon development. But about autonomous landing capability, there should be enough systems around now which can be used for this.

Even the robotic Dragon with full return and pin point landing capability, will make use of GPS and will land on a cement landing pad. Here we are talking about flying by inertia, recognizing the terrain, and deciding, with no external aid whatsoever (save the MRO maps), where there's a patch of terrain clean and leveled enough to actually land.
We are talking about 3D terrain recognition on real time. That's stuff that even the Google car is not ready for. There's a reason that Curiosity uses those huge wheels and suspension, and the Sky Crane works the way it does. Its to land on very unfriendly terrain.

The main issue with the Google car is dealing with other cars, moving stuff etc. 3D terrain recognition (and cars driving in it) in real time was demonstrated with the DARPA challenge many years ago, a more accurate comparison. As other people point out, ALHAT's been working on this for a while. I expect SpaceX to do better. But who knows, anything can be made to look good on paper.

[baldusi]

I have a few questions:
Are Dragon's legs rated for unleveled terrain?

As the paper plays with the numbers and features of the enhanced crew version of Dragon (with propulsive landing), your questions can only be answered from one inside Dragon development. But about autonomous landing capability, there should be enough systems around now which can be used for this.

Even the robotic Dragon with full return and pin point landing capability, will make use of GPS and will land on a cement landing pad. Here we are talking about flying by inertia, recognizing the terrain, and deciding, with no external aid whatsoever (save the MRO maps), where there's a patch of terrain clean and leveled enough to actually land.
We are talking about 3D terrain recognition on real time. That's stuff that even the Google car is not ready for. There's a reason that Curiosity uses those huge wheels and suspension, and the Sky Crane works the way it does. Its to land on very unfriendly terrain.

The main issue with the Google car is dealing with other cars, moving stuff etc. 3D terrain recognition (and cars driving in it) in real time was demonstrated with the DARPA challenge many years ago, a more accurate comparison. As other people point out, ALHAT's been working on this for a while. I expect SpaceX to do better. But who knows, anything can be made to look good on paper.

At which speed did the DARPA challenge cars run on the most difficult terrain? How far ahead did they had to plan ahead? What sort of data amount had to process, and how much time did they had?
Because it's not the same to analyze a couple of HD feeds a few meters ahead while you can stop an think a bit, against analyzing radar/lidar data plus some stereo images while literally falling down like a rock. The timing and the sort of decision making you have is very different.
In particular, at what height can you rally know that the terrain is smooth? I would guess you'd need at least 0.1m of resolution. That's a lot of data when you're falling like a rock. And once you've committed to a patch you're done.
Again, it's not that's not doable. It's that it might well be a couple of orders of magnitude more difficult than the DARPA challenge, and it would need some serious development. This is Mars we are talking about. Flying is not an option, just a soft crash landing.

[QuantumG]

Hmm.. Search did not reveal anyone posting this yet:

   http://digitalvideo.8m.net/SpaceX/RedDragon/karcz-red_dragon-nac-2011-10-29-1.pdf

Excuse the repetition if search has failed me. Not much new information in any case.

[ugordan]

Was posted here first, probably exactly where that came from: http://forum.nasaspaceflight.com/index.php?topic=26269.msg838572#msg838572

[QuantumG]

Thanks.. Search doesn't find attachment names

[ugordan]

Yes, without knowing exactly what to look for and then some, I find the search functionality here very difficult to tame. But we digress.