The need for a Moon direct *3* - ...continue here.

publiusr · 2005-09-09 14:10:10

TwinBeam wrote:

"Robot worshippers" is just a biasing epithet for people who want to get good science done without waiting for the nation to summon the will to pay for a manned Mars mission.

No--it is a statement of fact--lamenting how bomb-disposal robots and 100s length Thors that put two metric tonnnes less in orbit than 50 yr. old R-7 are helping to nickel and dime NASA to death.

Yet more Goldin era stagnation.

publiusr · 2005-09-09 14:10:38

TwinBeam wrote:

"Robot worshippers" is just a biasing epithet for people who want to get good science done without waiting for the nation to summon the will to pay for a manned Mars mission

No--it is a statement of fact--lamenting how bomb-disposal robots and 100s length Thors that put two metric tonnnes less in orbit than 50 yr. old R-7 are helping to nickel and dime NASA to death.

Yet more Goldin era stagnation.

Grypd · 2005-09-11 05:55:11

No Publiusr the Martian rovers and those robot probes are not the cause of NASA's cash flow problems. You can claim they are nickel and diming NASA but that is completely irrelevant compared to the likes of the ISS and the shuttles budget drain.

Robots so far have been an unqualified source of scientific and technical advances and are one of the few "sciences" that are making great leaps each day. Your decrying them as stagnation is simply so untrue. It is true that a geologist on Mars could have done all that the rovers have learned in a couple of days but we dont have people on Mars nor a capability to send them. Since we designed those rovers the science of robotics has made great leaps and the next generation of probes will be so much more.

It really comes down to capability versus time and of course cost and operating costs. This is why we are using robots and will have large contingents of them on the Moon where they will be able to operate a lot faster with Human oversight and do so much more. But they will not replace Humans, but they will support and empower human missions.

GCNRevenger · 2005-09-11 11:16:07

"the next generation of probes will be so much more."

No. The current MERs are already running up against the limitations of remote control, that is, they could actually do much more useful work if they weren't bound by the time lag from mission control. This probobly cannot be substantially fixed since computer software is still far too primative to do anything beyond drive on level terrain on their own reliably.

Sure you can bring more and better and bigger instruments, but this problem is not ever going to go away.

Grypd · 2005-09-11 14:45:00

The principle of the next generation of probes is that they will be more self controlled using techniques we have employed in the latest generation of UCAVs. Terrain avoidance with only general goals being given from Earth and then the robot taking the best way forward.

These can best be seen in the upcoming Darpa Grand Challenge.

GCNRevenger · 2005-09-11 17:02:30

The last DARPA challenge failed miserably

I think that experience with the MERs has clearly shown that the computer software to do what you are talking about just isn't there Grypd.

Grypd · 2005-09-12 16:16:09

But the last DARPA grand challenge was not a failure it brought a lot of money into research of advanced Robots. So what if no one actually managed to finish the challenge it was what they learned that made it so good. The next grand challenge is due very soon and there is a lot more money and resources now behind the teams. There is also a lot more teams too.

The MERS where designed almost ten years ago and as such are teleroboticaly controlled. They are also slow and need a lot of oversight. With luck this can be sorted in this the next generation and we get a decent set of hazard avoidance.

GCNRevenger · 2005-09-12 19:50:21

And they are supposed to make leaps and bounds in only one generation of robot-driven vehicles? And be reliable enough to risk a billion-dollar robot to not have any "hickup" with the advanced and highly complicated computers? Oh I doubt that, it will be years and years before we can trust a Mars rover to such a computer, since without a real-time link no babysitting is possible. Possible? Maybe. But the chances that it will be robust and reliable enough? Slim.

Oh, and don't forget, Mars has no GPS constellation

So you teach the thing to drive over, say, half the Martian terrain without active stop-and-go control? Swell, but thats only a portion of the time the rover will operate, it will still be stop-and-go when you reach your destination to actually do anything.

Robots are always going to lose in science-per-dollar

Grypd · 2005-09-13 17:41:27

And they are supposed to make leaps and bounds in only one generation of robot-driven vehicles? And be reliable enough to risk a billion-dollar robot to not have any "hickup" with the advanced and highly complicated computers? Oh I doubt that, it will be years and years before we can trust a Mars rover to such a computer, since without a real-time link no babysitting is possible. Possible? Maybe. But the chances that it will be robust and reliable enough? Slim.

But robotics are doing just that they are taking leaps and bounds. Though I have to agree that for the forseeable future Human oversight will be essential for a science based Martian/Lunar rover. Still that is due to us asking the rover to do a lot more and it needs human observation to say oh that is interesting. We will not be able to program in a curiosity that we can effectively use for scientific research. If we though task a robot with finding one or a limited numbers of elements then that robot will do it and will happily keep looking as long as it has power and is functional or till we tell it to stop. The Martian Rovers are just too universal things if we wish to use robots more effectively we have to have them specialise a bit. Tell a rover to do a mineral survey scanning for ilmenite and avoid hazards is a lot easier job than telling a robot to look for everything. The first is within our grasp and on October the 8th the grand challenge may show us this.

Oh, and don't forget, Mars has no GPS constellation

So build one we will have to anyway when we go to Mars in force as we will need a reasonable series of satelites and observation platforms to ensure communication and "weather" reports. (ps nothing to do with this question but Darpa will give a bonus to the team that completes the grand challenge without access to GPS- probably looking at the long view that and the need to fight a war when you have had your satelites knocked out or had to switch them of)

So you teach the thing to drive over, say, half the Martian terrain without active stop-and-go control? Swell, but thats only a portion of the time the rover will operate, it will still be stop-and-go when you reach your destination to actually do anything.

A robot is as only good as the purpose you want it to do. Currently we are using very small rovers which are doing a great job but are hardly getting a lot of the surface seen and are easily threatened by Dust and terrain hazards. If we wish to travel further afield then we need to send a bigger rover and preferably powered by a nuclear option battery. I have never said that a rover or robot is as good as a man doing the same job but we are not going to have people there for a while so we should wet our appetites with what the robots can do now. Any Mars mission will have a whole load of robots there to support the people anyway.

Tell a robot to look for something and it will very dillegently but it has to "know" what that thing looks like or is expected to look like and as such surprises are not really good for robots. It requires an extremely effective AI to be able to control a rover across an alien world and ignore the mundane and concentrate on the science. We are no where near that capability.

Robots are always going to lose in science-per-dollar

No robots win in the science for dollar route. What they lose in is universal flexibility and reasoning and more importantly the reason to go to space in the first place that of spreading mankind.

SpaceNut · 2005-09-15 06:01:35

A moon ride to remember

NASA researchers are doing a virtual moonwalk in the pastures surrounding Meteor Crater, negotiating cow pies in spacesuits as they dig for "samples" of planetary sediment and test vehicle prototypes destined for Mars and closer.
This is the eighth year the National Aeronautics and Space Administration has come to try out spacesuits, rover ideas and robots

Lots of extreme testing of concepts for lunar use designs are a must. Further enhancements of these items for Mars use will hopefully follow right away thou.

SpaceNut · 2005-09-16 10:47:04

Modification to a Previous NASA GSFC Notice: Propulsion Subsystem for the Lunar Reconnaissance Orbiter Project

Due to schedule constraints and cost savings, NASA/GSFC has determined to cancel our plans to issue Request for Proposal (RFP) NNG05121636R for the Propulsion Subsystem for the Lunar Reconnaissance Orbiter (LRO) Project office under the Robotic Lunar Exploration Program.

:?

publiusr · 2005-09-16 13:43:04

A robot is as only good as the purpose you want it to do.

A robot is only as good as its launch vehicle.

What kills me is how the robotics people will howl about how VSE will 'steal' money from science forgetting that the LVs we get out of the bargain will be superior to that Delta II crutch which forces people to drop science packages anyway due to mass constraints imposed upon them by Launch Vehicle development (or the lack of it) that the robots-only people have no interest in forwarding.

This was why The Planetary Society got chumped. They focused all their energy on advocacy and the payload--and got bit in the tush when the Volna flopped out on them not once but twice.

The Foton materials sat is based on the old Vostok hull which came as a result of human-centered space-flight--without which you don't get robot probes either. The new Venus craft is going atop the R-7--made by Korolov as more of a human-booster sold as an ICBM due to heavy warheads.

If it were up to the military--our largest rocket would be the Minuteman and the Russians' would have been Topol-M, neither of which could have launched Spirit and Opportunity.

We have focused too much on payloads at the expense of the rocket. It's time for the robots huggers to know their place. They might hate Griffin now--but when they have the Stick and HLLV--they will be grateful that they can kick that miserable underpowered Delta II to the curb.

Grypd · 2005-09-16 17:14:04

What kills me is how the robotics people will howl about how VSE will 'steal' money from science forgetting that the LVs we get out of the bargain will be superior to that Delta II crutch which forces people to drop science packages anyway due to mass constraints imposed upon them by Launch Vehicle development (or the lack of it) that the robots-only people have no interest in forwarding.

What many people howled about had nothing to do with Robots. NASA accumalated a lot of things that came under a science umbrella. Look at all the research that NASA did for Aeronautics and even in computer research. It also did studies in material science and food crops anything that could be called space research NASA was to pay for. This is why NASA's budget was so stretched.

NASA has a lot of science satelites from the HUBBLE down to the probes operating to look at the weather and Ozone depletion. NASA to do the VSE will have to cut a lot of money from all of these. HUBBLE is already doomed but so is replacement ozone and Earth research probes.

If you really look at the VSE it actually increases the money that NASA will spend on Robotics. It will use the Robonaut for the first time and NASA will have to actually start manufacturing them. There is also a lot more money for what is called core research this is the exploration probes to the Moon and Mars and a lot will have robotic elements. We are to see a whole new generation of rovers.

Dont worry publiusr the robotics fanatics are more than pleased with the VSE you see their research has for a long time been considered one of the weakest disciplines and money is often cut or have been given shoestring budgets in the first place of course most Roboticists are engineers and not scientists that probably explains it.

If it were up to the military--our largest rocket would be the Minuteman and the Russians' would have been Topol-M, neither of which could have launched Spirit and Opportunity.
We have focused too much on payloads at the expense of the rocket.

We did ignore the rockets as it really does not matter what they did as in the end it is the payload that matters. What is a rocket but a means to an end that of delivering a payload.

But since the glory apollo days we have been stuck in LEO and we did not need a generation of superpowerful rockets. NASA said we had the shuttle to do everything. So we went nowhere fast.

It's time for the robots huggers to know their place. They might hate Griffin now--but when they have the Stick and HLLV--they will be grateful that they can kick that miserable underpowered Delta II to the curb.

You may dislike robots and despise those who work on them but then you are so like the old NASA, that Griffin is so determined to replace. And I am so glad that with VSE there is the chance for more roboticists to get further than the workshop. You may call me a robot hugger and actually I smile at that reference then again I know what im talking about.

And Delta wont go away we will still need to send satelites up and for that Delta is perfect. We just dont need to mix the Human spaceflight and satelites launches together to give the appearance of some form of cost savings.

Commodore · 2005-09-19 15:19:13

Robotic probes still have a long way to go before they reach the AI barrier. A nuclear powered probe or lander could get far more done than anything we've fielded thus far.

But as always, probe are useless unless we apply what they tell us to manned flight.

SpaceNut · 2005-09-20 04:51:46

Robotic probes still have a long way to go before they reach the AI barrier. A nuclear powered probe or lander could get far more done than anything we've fielded thus far.

Would a system simular to the one on probes like voyager or perhaps casini weigh to much?

But as always, probe are useless unless we apply what they tell us to manned flight.

Which makes me would what is up with the LRO mission. Are the just delaying the timeline for when it is needed as in sometime closer to just before the manned mission of 2018.

GCNRevenger · 2005-09-20 12:30:23

"If we though task a robot with finding one or a limited numbers of elements then that robot will do it and will happily keep looking as long as it has power and is functional or till we tell it to stop"

You can't even program a robot to do that yet, the software is still far, far, far too primative. How can you teach a robot to look for things that even trained geologists have a rough time spotting? Or following terrain trends? No, complete human oversight for all operations besides transit over easy terrain will be nessesarry. This is why humans will always, always win in the end over a "dumb" rover.

"The first is within our grasp and on October the 8th the grand challenge may show us this."

And what does the Grand Challenge software do? All it does is tell the rover not to drive directly into an obsticle. Thats all. It has no context whatsoever of what it is "looking" at, all it knows is the image in front of it isn't unobstructed ground.

A Martian GPS network would have to consist of like a dozen or more units to provide decent decimeter-scale accuracy. That is expensive, and humans wouldn't need it.

"Tell a robot to look for something and it will very dillegently but it has to "know" what that thing looks like or is expected to look like and as such surprises are not really good for robots. It requires an extremely effective AI to be able to control a rover across an alien world and ignore the mundane and concentrate on the science. We are no where near that capability."

Which is why humans win big, dollar-for-dollar, on the quantity and quality of science versus robot rovers.

SpaceNut · 2005-09-20 20:03:20

Sort of fits nicely with robotics needing to
Digging "Moon Dirt" is NASA's Fifth Centennial Challenge
And needing to do so with little human intevention.

NASA today announced the Regolith Excavation Challenge, a new Centennial Challenges prize competition that will award $250,000 to the winning team and has the potential to significantly contribute to the nation's space exploration goals. The competition is in collaboration with the California Space Education and Workforce Institute (CSEWI).
The Regolith Excavation Challenge will award the prize money to the team that can design and build autonomously operating systems to excavate lunar regolith, or "moon dirt," and deliver it to a collector.

Grypd · 2005-09-21 15:25:52

"If we though task a robot with finding one or a limited numbers of elements then that robot will do it and will happily keep looking as long as it has power and is functional or till we tell it to stop"
You can't even program a robot to do that yet, the software is still far, far, far too primative. How can you teach a robot to look for things that even trained geologists have a rough time spotting? Or following terrain trends? No, complete human oversight for all operations besides transit over easy terrain will be nessesarry. This is why humans will always, always win in the end over a "dumb" rover.

What do geologists do to search for minerals they use small explosives move a bit and listen to the reflections of sound of the explosions. They use ultra sonics to detect what the soil content is. They also use a good eye to look around and in space this means taking a lot of pictures to see what can be found. A lot of this can be automated with more promising sites warranting further investigation by human contact.

"The first is within our grasp and on October the 8th the grand challenge may show us this."
And what does the Grand Challenge software do? All it does is tell the rover not to drive directly into an obsticle. Thats all. It has no context whatsoever of what it is "looking" at, all it knows is the image in front of it isn't unobstructed ground.

Terrain observation is a good science for robotics it consists of a robot being able to choose the best means to go around and actually what the terrain means. All very good science to learn.

A Martian GPS network would have to consist of like a dozen or more units to provide decent decimeter-scale accuracy. That is expensive, and humans wouldn't need it.

But if we start to send a lot of people we will need them desperatly and the plan to create a Martian internet is an example of something that actually increases capability and any such coverage also can be used for general GPS signals. Not likely for a while but still it will be needed.

"Tell a robot to look for something and it will very dillegently but it has to "know" what that thing looks like or is expected to look like and as such surprises are not really good for robots. It requires an extremely effective AI to be able to control a rover across an alien world and ignore the mundane and concentrate on the science. We are no where near that capability."
Which is why humans win big, dollar-for-dollar, on the quantity and quality of science versus robot rovers.

Granted, but it is the ability to stay on station and keep working and allowing research to continue that makes them interesting. Again there is also the lightness of the devices that does give them advantages.

GCNRevenger · 2005-09-21 16:45:59

I don't think that robots are even smart enough to do that yet...

Take pictures of what? The robot doesn't know what to focus on or look for, so it would be totally random.

And you want it to be smart enough to use explosives or carefully bury highly sensitive siesmographs? Or drill into promising sites and take samples? I don't think so.

No, no it can't be automated. We aren't anywhere near that good yet. Terrain avoidance is also in its infancy, this is something we can't do reliably enough yet to risk a $0.5-1.0Bn+ robot, and probobly aren't for some years.

The usefulness of robots without direct, nearly real-time human oversite is simply very limited.

Grypd · 2005-09-21 17:15:10

We are not asking them to do interpretation just standard surveys. Deploying an explosive charge uniformily is something that robots can do now very well look at the wheelbarrow class of land based anti ordinance robots or the use of UAVs which are completely unmanned and out of Human control. Especially the air force ones that are being used to bomb targets.

Recording what they pick up is impostant as well as quite a boring monotonous job.

Satelites from orbit will give us the place to check and short of sending people every time we can send a robot and it will simply photograph everything from standard reference points. These then can use Human asstes better in deciding what and how further investigation should proceed. If the Initial survey is promising then we can teleoperate the rover from the base and further research can be done.

The most boring part and one that we will not have the manpower for is to drive a robot to a site for its mission to start. We do not have the time or manpower to control each robot that is needed. If the robot can get itself there then we are saving prescious human time. This especially important for Missions to Mars but not necessarily so for the Moon

GCNRevenger · 2005-09-21 18:15:06

I'm saying that they can't even do a "standard survey" efficently because their search will be, though methodical, will be random and mindless.

Most mine-clearing robots are still operated by remote control, and the automatic ones don't work over rough terrain.

Flying UAVs have the advantage of not having anything to fly into that isn't on a GPS map.

Satelites can only provide views from directly overhead reliably, and thus will probobly miss lots of interesting objects. They also don't have very good resolution, and can't see anything smaller then a meter or two across.

Robots can't be trusted to navigate anything except smooth, unobstructed terrain reliably, the risk to such a huge investment is too big. Much of the time a robot will be operating it will be on site, not transiting anyway.

Grypd · 2005-09-22 16:33:11

Hi GCN

That may be true right now but 5 years down the line :?

Here is NASA jumping on the Grand Challenge bandwagon but this one with robots mining and collecting Lunar regolith to be processed or used in construction materials this to happen without human oversight.

NASA challenge to roboticists to build autonomous Moon diggers

Incidentally our mapping of Mars and knowledge of its terrain is a lot better than what we currently have of the Moon.

SpaceNut · 2005-09-27 09:36:22

NASA to Show Intelligent Space Robots in Action at Ames 'Marscape'

MEDIA ADVISORY: You are invited to the outdoor 'Marscape' at NASA Ames Research Center, Moffett Field, Calif., Monday, Oct. 3, from 11 a.m. to 2 p.m. PDT to view prototype space robots in action. Reporters will be able to interview NASA researchers about the artificial intelligence that may some day help robots and human beings on Earth and Mars work together.

The two robots – 'K-9' and 'Gromit' – are smart enough to make decisions about how to achieve objectives on a planet or moon without detailed instructions from human beings. Researchers will also demonstrate 'mobile agent' software that may someday help robots and human beings on Earth, the moon and Mars communicate with each another.

Gee I guess we have come a few steps further..

SpaceNut · 2005-09-27 12:51:30

Another reference article to the same:
NASA to Show Intelligent Space Robots in Action at Ames Marscape

SpaceNut · 2005-09-30 19:09:30

Nasa Decision for lunar robotic exploration missions to come.

NASA Selects Team to Build Lunar Lander

NASA's Deputy Associate Administrator for the Exploration Systems Mission Directorate Doug Cooke today announced the selection of NASA's Marshall Space Flight Center, Huntsville, Ala., and Goddard Space Flight Center, Greenbelt, Md., to lead a team in the development of a lunar lander spacecraft.

This lander is to provide information that will not becoming from the LRO mission of 2008.

The lander is tentatively planned for launch as early as 2010. It will demonstrate the ability for precision landings at targeted locations on the moon; evaluate landing zone environment; and determine if lunar resources can support a sustained human presence.

Reasons for needing lander happen to be to confirm water:

This mission will have as a primary objective to determine whether there is water-ice in the permanently dark areas within craters in the moon's polar regions. The existence of water-ice has important implications in living off the land when we return with human explorers

New Mars Forums

Announcement

#101 2005-09-09 14:10:10

Re: The need for a Moon direct *3* - ...continue here.

#102 2005-09-09 14:10:38

Re: The need for a Moon direct *3* - ...continue here.

#103 2005-09-11 05:55:11

Re: The need for a Moon direct *3* - ...continue here.

#104 2005-09-11 11:16:07

Re: The need for a Moon direct *3* - ...continue here.

#105 2005-09-11 14:45:00

Re: The need for a Moon direct *3* - ...continue here.

#106 2005-09-11 17:02:30

Re: The need for a Moon direct *3* - ...continue here.

#107 2005-09-12 16:16:09

Re: The need for a Moon direct *3* - ...continue here.

#108 2005-09-12 19:50:21

Re: The need for a Moon direct *3* - ...continue here.

#109 2005-09-13 17:41:27

Re: The need for a Moon direct *3* - ...continue here.

#110 2005-09-15 06:01:35

Re: The need for a Moon direct *3* - ...continue here.

#111 2005-09-16 10:47:04

Re: The need for a Moon direct *3* - ...continue here.

#112 2005-09-16 13:43:04

Re: The need for a Moon direct *3* - ...continue here.

#113 2005-09-16 17:14:04

Re: The need for a Moon direct *3* - ...continue here.

#114 2005-09-19 15:19:13

Re: The need for a Moon direct *3* - ...continue here.

#115 2005-09-20 04:51:46

Re: The need for a Moon direct *3* - ...continue here.

#116 2005-09-20 12:30:23

Re: The need for a Moon direct *3* - ...continue here.

#117 2005-09-20 20:03:20

Re: The need for a Moon direct *3* - ...continue here.

#118 2005-09-21 15:25:52

Re: The need for a Moon direct *3* - ...continue here.

#119 2005-09-21 16:45:59

Re: The need for a Moon direct *3* - ...continue here.

#120 2005-09-21 17:15:10

Re: The need for a Moon direct *3* - ...continue here.

#121 2005-09-21 18:15:06

Re: The need for a Moon direct *3* - ...continue here.

#122 2005-09-22 16:33:11

Re: The need for a Moon direct *3* - ...continue here.

#123 2005-09-27 09:36:22

Re: The need for a Moon direct *3* - ...continue here.

#124 2005-09-27 12:51:30

Re: The need for a Moon direct *3* - ...continue here.

#125 2005-09-30 19:09:30

Re: The need for a Moon direct *3* - ...continue here.

Board footer

Re: The need for a Moon direct 3 - ...continue here.

Re: The need for a Moon direct 3 - ...continue here.

Re: The need for a Moon direct 3 - ...continue here.

Re: The need for a Moon direct 3 - ...continue here.

Re: The need for a Moon direct 3 - ...continue here.

Re: The need for a Moon direct 3 - ...continue here.

Re: The need for a Moon direct 3 - ...continue here.

Re: The need for a Moon direct 3 - ...continue here.

Re: The need for a Moon direct 3 - ...continue here.

Re: The need for a Moon direct 3 - ...continue here.

Re: The need for a Moon direct 3 - ...continue here.

Re: The need for a Moon direct 3 - ...continue here.

Re: The need for a Moon direct 3 - ...continue here.

Re: The need for a Moon direct 3 - ...continue here.

Re: The need for a Moon direct 3 - ...continue here.

Re: The need for a Moon direct 3 - ...continue here.

Re: The need for a Moon direct 3 - ...continue here.

Re: The need for a Moon direct 3 - ...continue here.

Re: The need for a Moon direct 3 - ...continue here.

Re: The need for a Moon direct 3 - ...continue here.

Re: The need for a Moon direct 3 - ...continue here.

Re: The need for a Moon direct 3 - ...continue here.

Re: The need for a Moon direct 3 - ...continue here.

Re: The need for a Moon direct 3 - ...continue here.

Re: The need for a Moon direct 3 - ...continue here.