Author Topic: NASA Robotics and Artificial Intelligence Directorate?  (Read 3183 times)

Offline Darkseraph

  • Full Member
  • ****
  • Posts: 711
  • Liked: 475
  • Likes Given: 152
By the late 2030s, chemical rocketry will have improved marginally in performance, if at all. And reusability might make things cheaper, but certainly not aircraft like costs or operations. Just better than now. The round trip journey to Mars will still be a minimum of more than a year and will likely have significant dangers and very costly.

Not bringing humans back to cut some of the costs is not a near term option for preliminary science missions. That's a colonization effort, which NASA has zero mandate for and something we will do when we know it is feasible for that planet to sustainably host human civilization. NASA is supposedly targeting the 2030s for Mars missions.

By the 2030s, Robotics and AI will have very likely advanced far more than rocketry in the same time period. There is an idea that we should send humans to Mars orbit in the 2030s to remote control drones on the surface. But this all got me thinking, that just keeping humans in orbit to avoid some lag is dangerous without the rewards of a landing, but still pretty expensive. What if you used AI instead to overcome the lag, and sent basically androids/intelligent autonomous robots to the surface of Mars?


So, I know this will go down like a lead upperstage on a human space flight oriented forum. But I was thinking, would it be a terrible idea to have AI&Robotics as a separate NASA directorate, tasked with both incorporating new technology as well as developing it for the end purpose of creating semi-autonmous  teams of robots to explore planetary surfaces? They might even been humanoid, engaging the public in a way a rover like thing might not. But you wouldn't need to bring them back, and going by amazing length of the MER rovers, they could stay on the surface doing more science for far longer than any human mission. If you sent several together, like with Astronauts, there is even the possibility they could repair each other. They could explore areas out of line of sight, such as Lava tubes and underground areas.

The economic and social payback of such an idea is really obvious and down to Earth. Helping the advancement of such systems for exploration of moons and planets would be a huge boon on Earth and its industries. Research into complex movement, machine learning, hazard avoidance, self repair and a whole bunch of fields would need to be carried out. And this would focus all those into an integrated whole.


So, without shooting me...what do people think of that?
"For a successful technology, reality must take precedence over public relations, for Nature cannot be fooled." R.P.Feynman

Offline RonM

  • Senior Member
  • *****
  • Posts: 3340
  • Atlanta, Georgia USA
  • Liked: 2231
  • Likes Given: 1584
Re: NASA Robotics and Artificial Intelligence Directorate?
« Reply #1 on: 04/04/2015 08:07 pm »
I think the timing is off. AI capable of making decisions on the level of a trained scientist or engineer probably will not be available by the 2030s. That is far future, if it is even possible.

Sophisticated robotics controlled by astronauts from Mars orbit might be good enough to delay manned landings until colonization begins.
« Last Edit: 04/04/2015 08:09 pm by RonM »

Offline Darkseraph

  • Full Member
  • ****
  • Posts: 711
  • Liked: 475
  • Likes Given: 152
Re: NASA Robotics and Artificial Intelligence Directorate?
« Reply #2 on: 04/04/2015 08:45 pm »
I don't mean human level AI. That is possibly a hundred years away if not more.

Just far better AI with advanced robotics that you can continuously update, repair and so on.. Humans controlling robots from Mars orbit is..not worth the risk to their lives, in my opinion for the return in science. It doesn't advance the state of the art very much either. I don't know why that mission keeps coming up, it solves so few of the problems of sending humans to Mars, but has almost all of the risks.  :/
"For a successful technology, reality must take precedence over public relations, for Nature cannot be fooled." R.P.Feynman

Offline RonM

  • Senior Member
  • *****
  • Posts: 3340
  • Atlanta, Georgia USA
  • Liked: 2231
  • Likes Given: 1584
Re: NASA Robotics and Artificial Intelligence Directorate?
« Reply #3 on: 04/04/2015 09:19 pm »
I don't mean human level AI. That is possibly a hundred years away if not more.

Just far better AI with advanced robotics that you can continuously update, repair and so on.. Humans controlling robots from Mars orbit is..not worth the risk to their lives, in my opinion for the return in science. It doesn't advance the state of the art very much either. I don't know why that mission keeps coming up, it solves so few of the problems of sending humans to Mars, but has almost all of the risks.  :/
Getting humans to Mars orbit is half the battle. The other half is landing and taking off. NASA Mars mission planning has spacecraft going into Mars orbit before landing. The purpose of controlling robots from Mars orbit is to have something to do while you're testing the mission architecture before sending a lander on the next mission.

Of course, if you're using what people think the SpaceX MCT architecture is about, then it's going for landing from the start. No need to control robots from orbit because you're not messing around with orbit. You're coming in hot.

With better AI, but nowhere near human level, robots would be more independent. We already see the beginnings of that with Curiosity. It can do some of the driving by itself. In the 2030s, robots will be more independent and not need as many instructions from Earth.

Offline Paul451

  • Senior Member
  • *****
  • Posts: 3553
  • Australia
  • Liked: 2518
  • Likes Given: 2180
Re: NASA Robotics and Artificial Intelligence Directorate?
« Reply #4 on: 04/05/2015 03:46 am »
In-space robotics seems to lag well behind commercial robotics and expert systems, which lags behind lab-level robotics and AI.

So I'd say that the level of automation available for Mars in 2030 would be roughly equivalent to a more reliable version of cutting edge development in AI/robotics labs today.
« Last Edit: 04/05/2015 03:48 am by Paul451 »

Offline Oli

  • Senior Member
  • *****
  • Posts: 2467
  • Liked: 605
  • Likes Given: 60
Re: NASA Robotics and Artificial Intelligence Directorate?
« Reply #5 on: 04/05/2015 05:35 am »
The round trip journey to Mars will still be a minimum of more than a year and will likely have significant dangers and very costly.

But I was thinking, would it be a terrible idea to have AI&Robotics as a separate NASA directorate, tasked with both incorporating new technology as well as developing it for the end purpose of creating semi-autonmous  teams of robots to explore planetary surfaces?

- Let's not forget that Mars is far from being the only interesting exploration target in the solar system. With human spaceflight you're realistically limited to exploring the Moon, some nearby asteroids and Mars. In that sense it's a dead end "technology".

- Well NASA has a planetary science division, maybe it belongs there.

Offline Darkseraph

  • Full Member
  • ****
  • Posts: 711
  • Liked: 475
  • Likes Given: 152
Re: NASA Robotics and Artificial Intelligence Directorate?
« Reply #6 on: 04/05/2015 04:11 pm »
The round trip journey to Mars will still be a minimum of more than a year and will likely have significant dangers and very costly.

But I was thinking, would it be a terrible idea to have AI&Robotics as a separate NASA directorate, tasked with both incorporating new technology as well as developing it for the end purpose of creating semi-autonmous  teams of robots to explore planetary surfaces?

- Let's not forget that Mars is far from being the only interesting exploration target in the solar system. With human spaceflight you're realistically limited to exploring the Moon, some nearby asteroids and Mars. In that sense it's a dead end "technology".

- Well NASA has a planetary science division, maybe it belongs there.

Indeed, Mars is not the only interesting target..but that increases the attractiveness of robotic systems since further out environments have even more lag, potentially worse radiation issues/gravity. Robots don't get lonely or have psychological issues as of yet. You can put them into ehh "hypersleep"..basically low powered mode. We are not sure if such a thing is possible with humans.

I know the tech for this lags as a poster above mentioned, but that's exactly why I thought this would be an interesting directorate, to really jump start the effort and work with actors in the private sector, military and universities. So not only doing research and development for new technology that gets transferred to the private sector, but also buying some of these services from companies that are doing their own research, such as IBM, Google, Boston Dynamics and so on.

I think space robotics seems to be something the military is taking up more tbh under DARPA and the Air Force. Things like the X37B, Orbital Express, and so on.

What strikes me most about sending robots is for the Mass needed just to land humans one way, you could send an incredible amount of robotic systems and get far more science return. The objection I've seen Zubrin and a few others put towards this is that a field geologist could have done in days what MER rovers did in years.....if we used about 50 times the mass....we could just use the same amount of mass to send more sensors, higher data rate sats in Mars orbit, rovers, sophisticated robots, refueling and recharging capabilities, self repair mechanisms and so on. You could send a lot of stuff for the money and mass require even to send humans to just Mars orbit. And you don't have to wait till the 2030s to initiate it, you can progressively send more advanced missions every two years, which have more mass, test new things, visit new areas and so on. The current program of record calls for a 2020 rover with no real plans until some time in the mid 2030s humans might land. That seems like a bad idea, to have a gap like that.
"For a successful technology, reality must take precedence over public relations, for Nature cannot be fooled." R.P.Feynman

Offline Port

  • Full Member
  • *
  • Posts: 111
  • Germany
  • Liked: 13
  • Likes Given: 8
Re: NASA Robotics and Artificial Intelligence Directorate?
« Reply #7 on: 04/05/2015 04:37 pm »
I think the timing is off. AI capable of making decisions on the level of a trained scientist or engineer probably will not be available by the 2030s. That is far future, if it is even possible.

Sophisticated robotics controlled by astronauts from Mars orbit might be good enough to delay manned landings until colonization begins.

you should really read this:

http://waitbutwhy.com/2015/01/artificial-intelligence-revolution-1.html
http://waitbutwhy.com/2015/01/artificial-intelligence-revolution-2.html

long posts, but worth every laugh and thought - also very informative
the best way to describe things that will happen in regard to AI (from the thread) is a picture i attached

it is actually the reason why Hawking, Musk and other people in this field are very worried
« Last Edit: 04/05/2015 04:38 pm by Port »

Offline Darkseraph

  • Full Member
  • ****
  • Posts: 711
  • Liked: 475
  • Likes Given: 152
Re: NASA Robotics and Artificial Intelligence Directorate?
« Reply #8 on: 04/05/2015 06:54 pm »
Actually I think that kind of thinking promoted by Kurzweil and others Singulatarians is on the unscientific woo end of the spectrum, that simple increases in computing power are not magical pixie dust and won't result in general artificial intelligence by themselves. And computers at the moment are still really dumb and make unintelligent mistakes. Humans for all their stupidity, make intelligent errors, that show they a degree of insight, that they comprehend what they're being asked to do.


But I do believe they can be very useful systems to develop for exploration of other planets and moons and there is a lot of room to improve them between now and the 2030s.  They can be far more autonomous and upgraded as the geologists, biologists and other experts on the ground send them either new tools or new software to advance their investigations. Ultimately, if colonization of space is to be achieved, I think it will require a lot of these systems, preparing for human settlers and going to environments that are too risky for humans.
"For a successful technology, reality must take precedence over public relations, for Nature cannot be fooled." R.P.Feynman

Offline Jim

  • Night Gator
  • Senior Member
  • *****
  • Posts: 37440
  • Cape Canaveral Spaceport
  • Liked: 21451
  • Likes Given: 428
Re: NASA Robotics and Artificial Intelligence Directorate?
« Reply #9 on: 04/05/2015 07:42 pm »
Why?  There wasn't a rover directorate and they got developed.  Automated spacecraft get developed all the time, but there is no specialized directorate. 

Tags:
 

Advertisement NovaTech
Advertisement Northrop Grumman
Advertisement
Advertisement Margaritaville Beach Resort South Padre Island
Advertisement Brady Kenniston
Advertisement NextSpaceflight
Advertisement Nathan Barker Photography
1