Follow Slashdot blog updates by subscribing to our blog RSS feed

 


Forgot your password?
Close
typodupeerror
×
Science

NASA and AI Testing 113

Posted by Hemos from the check-out-the-neat-stuff dept.
NapalmCheese writes "NASA tested their AI (Remote Agent) in space last May and it passed with flying colors. The articles makes it sound like HAL, I don't know if that is a good thing, but definately cool. The first article is found at The JPL and is nice and informative, the second article found at, rax is even more so. Hmmmm, if only it had a big red eye. "
This discussion has been archived. No new comments can be posted.

NASA and AI Testing More

NASA and AI Testing

Comments Filter:
  • How about if, during the 5 minute lag between Ground Control and the Mars Explorer, the rover had managed to knock itself upside down. A person could easily nudge it back over with their foot, but radio transmissions from Earth aren't going to do squat. Also, that 5 minute lag only gets longer and longer the further away from Earth you go, so having a person on the scene will dramatically affect the outcome in time-sensitive, mission-override type scenarios. AI is great for accomplishing the tasks the engineers can think of, but not for the unknown.
  • I'm working on two different experiments scheduled to be installed on the space station. I'll tell you that my motto is that the space station is for human (biological) science and not physical science. There are so many problems with trying to make everyone's experiments co-exist; it's a nightmare.

    I support human spaceflight and think it is very necessary. We are learning a considerable amount by researching biology in a new evironment (with microgravity). That's how science is supposed to work. There was a great program on the Discovery Channel last night (The Invisible Force, check it out) that addressed this.

    However, somehow during the politics, the space station was viewed as the end-all solution to putting experiments into space to support all of the funding. After all, the more customers you have, the more support you get in Congress. Now it's this big complex structure that's made it even more expensive.

    I wish it was scaled back for the sole purpose of biological/human research and not this complicated experimental facilicity that's supposed to satisfy everyone. Then more money would be available to experimenters for their research.

    ~afniv
    "Man könnte froh sein, wenn die Luft so rein wäre wie das Bier"
  • Is this really an AI? Aren't there a set of laws that define what is truly an AI and what is just Agent software on a par with Amazon.com's ability to pick out books for me based on my past selections?
  • From the comments on the rax page it seems like they're using a Bayesian network to diagnose errors and identifying the causes of the errors. If so we should thank Microsoft in part for hiring some of their researchers.

    It's surprising but Microsoft Research [microsoft.com] has a bunch of researchers that have done a lot of important research in Bayesian networks and AI. For example, David Heckerman developed PATHFINDER for his Ph.D. thesis. (PATHFINDER is a program that helps to diagnose lymph node diseases, the last version PATHFINDER IV was just as good or better than most pathologists). In fact, I think one of them wrote a paper on diagnosing and troubleshooting PC problems which may have helped the people at JPL write the MIR software.

  • > We could send a few dozen people to the moon,
    > but what good is that going to do us? No, we're > better off waiting till we have something
    > worthwhile, like at least go to Mars.

    We have to learn to walk before we can fly, so to speak. The moon offers us a number of ways to test the technologies that we would use in getting to and staying on Mars for extended periods.
  • This project also had ion engines as a propulusion unit. In fact, their was a software error that messed up the timing of the ion engine components so it would shut off during this flight. The software was able to work around it automatically though.
  • There is a certain undeniable problem about life on earth - every species is NOT sustainable. Humans are not going to figure out a way to decrease the human population, so eventually we will run out of resources and there will be mass starvation, etc. That is, unless another species wipes us out first - Insects anyone? Long after humans (and other animals) have disappeared from earth, the insects will still thrive.

    So, if we want to ensure that the human species, and all the other animal species on earth survive far into the future, it is our RESPONSIBILITY to explore space, and find somewhere else to settle. The space station is but one important part of that space exploration. No other species on earth has the ability to seek out new worlds, and if we as humans don't do it, we are sealing the fate of every living thing on the planet (except for maybe the dominant species, those darned insects). That wouldn't be very responsible, would it?

    One can think of that neat story about a guy named Noah and his ark, where he took all the animals of the earth to save them from the flood. We will eventually be flooded by overpopulation, and we'll be in need of an ark.
  • What is the point of a space station if it is right outside the earth's atmosphere? Lets be serious here, we should colonize the moon first...
    Here's why:
    1) We are running out of space on earth.
    2) Rockets can landon the moon to refuel if they had to.
    3) The low levels of gravity would make for awesome basketball games (then i might actually make the team...)
    4) If done right, (artificial light, heat, air (like a space station) it could be an even better International Space project, because people from different countries would actually want to go there.
  • I found these through a search using google.com:

    for Allegro Common Lisp: www.neci.nj.nec.com/homepages/mann/software-availa ble.html

    for GNU Common Lisp:
    stat.umn.edu/~luke/www.html
  • The space station is just one step. Once we have a space station, it will be tons easier doing other things in space. We can really explore the near solar system.

    Also, why not fund oceanic research? Instead of transfering funds form one scientific endeaver to another, lets take some money out of that huge DoD budget. It is big enough as it is.

    --

  • You can't just drop the space station though. ROI is great....but you have to keep public support high, or there will be zero funding for space exploration.

    The only way to keep public support high is to treat space as The Final Frontier and to explore it with people. It's sad (I agree with you, these probes could teach us more than the space station) but a fact of life none the less.
    Oh well.

  • Re:Drop the Space station, fund this kind of stuff (Score:5)

    by gavinhall ( 33 ) on Tuesday June 08, 1999 @02:47AM (#1862316)
    Posted by ionic:

    I disagree whole heartedly! I both think and feel that sending Man into Space is paramount; I mean think about it this way Earth is the only stomping ground of our civilization, if Earth goes we go. Therefore, the more that we push sending Man into Space and more directly seeding Mankind on the Moon, and planets like Mars ensures that should something happen to one locale in Space then at least something of Mankind would survive. The problem today is that the "leaders" in governmentally pushed space exploration (the USA) cannot get its populace behind spending more money on manned or any other type of Space exploration. Think about it, how long did it take to get the Space station in place, and we continue to sit on our "asses" looking at the moon, Mars and the asteroid belt, dreaming about them and doing nothing to bring them to reality (don't get me wrong there are some porojects whish are aimed at this but they are few and far between). If you want to talk about economics well, my guess is that there are plenty of precious resources within the asteroids which are waiting to be mined even as we jabber about the econonic feasibility of Space ventures (be it govermentally or privately funded). To get back to my major point a whole lot of us need to awaken to the fact that if there is a push in the direction of Space it will sprout new industries to support itself thereby rendering the enconomic issue a moot point. And as a side benefit of making local Space no longer a frontier we spread the seeds of Mankind onto multiple types os soil where it can both grow and wither, but we get out there.

    Mike Hay
  • I guess I have seen T2 to many times! An evil AI coudl wipe us all out of existance.
  • > I believe that artificial intelligence is
    > defined as "technology required to perform
    > impossible computing tasks"

    I think that AI is commonly defined using Alan Turing's definition (late fourties). AI is when you can no longer reliably determine that a talk session is with the computer AI or a real human. (Transposed into modern terminology).

    This is definitively not an AI: it just has a set of objectives, and attempts to achieve them knowing that it has reduced abilities. (Or by power cycling the subsystem to see if it comes back to life).

    If you analyse the log, you can see that it takes RAX over an hour to deduce that it can operate the craft with the missing attitude jet.

  • An "expert system" could accomplish the tasks of their agent, as could a genuinely "intelligent" agent. The real answer as to whether it is AI or not has to do with the design of the code.
  • Check out comp.ai for more information on the satelite. Some of the people working on the project having been posting updates and would probably be willing to answer questions.
  • I can't disagree with you more. Space travel is NOTORIOUSLY inefficient--it costs millions upon millions of dollars per component on each shuttle and probe NASA and other space agencies send into orbit. I should know, I work for Lockheed Martin. Absolutely nothing within the current world socioeconomic and scientific structure suggests that space travel is going to be feasible on a large scale any time in the year future--the space program has been active for decades and the price per mission keeps going up, not down.

    You say: "We continue to sit on our 'asses' looking at the moon, Mars, and the asteroid belt, dreaming about them and doing nothing to bring them back to reality."

    We made a manned moon landing in 1969, 30 years ago. It's a really big rock. There's nothing there. If there was, we'd still be going there.

    We've been to Mars, via several unmanned probes. What's the point of sending people there? Again, it's just a big rock. If there was anything interesting, our super-expensive, super-accurate probes and surveying equipment would have picked it up by now.

    We know what asteroids are made of. Scientists have been studying impact sites of landed meteorites for years. Guess what? They're rocks. Metallic rocks perhaps, but certainly not rare enough to justify SPENDING millions upon millions of dollars to retrieve them.

    You claim that "if there is a push in the direction of Space it will sprout new industries to support itself thereby rendering the enconomic issue a moot point." Why is this? Things aren't going to get less expensive just because we have private companies footing the bill instead of NASA. Besides, there's still nowhere to go, and no point in going there.

    Your romantic notions are, well, romantic, but they have absolutely no scientific basis behind them.
  • Well, in T2, the AI wasn't what I would define as "evil". But in The Matrix, when the one agent began to relate how he hated "this place", and how humans just stink, and their stink was all over the place, and how he hated them. That fits my definition of evil. When the machine ceases to be cold, logical, reasonable, and began to hate.

    "The number of suckers born each minute doubles every 18 months."
    -jafac's law
  • NASA has a long history of using Lisp for such things: In fact, they programmed the Mars Pathfinder robot [franz.com] in Lisp (using Allegro CL).

    Full disclosure: I'm doing contract PR work for Franz. Stories like this are like candy to me. :) --Tom pr@franz.com

  • I believe that mankind knows little about the climate on a geological scale and can not yet conclusively prove we are not in danger of a catastrophic climate change caused by other than man-made circumstances.

    I believe a comment by Nobel Prize winning chemist Kary Mullis that if one were to do actuary tables on the most pressing threats to mankind, one would find that a "Extinction Level Event" caused by an asteroid is possible, and our politicians will not significantly address this problem until it's to late because of financial concerns.

    I believe that NASA has not focused it's efforts correctly in these matters. I believe NASA has focused on "Star Wars" type programs (both in a Cold War, Ronald Regan sense, and in a George Lucus glamorous space travel sense), when in fact they should focus more on conservation concerns such as meteor impacts.

    I believe Space Studies are disproportionately funded to Under Sea studies. Hollywood has glamorized space studies, and under played deep sea exploration. Our children are encouraged to "reach for the stars" by Hollywood, and, I believe we already know more about the stars than we do about our own planet (specifically, what's going on under the seas.

    I believe the answers to many of our social and cultural problems can be found through science, and science has been given a bad hand to play by the media and government. Under Sea exploration and studies my solve countless problems in geological understanding, food and fuel resources, and many other significantly neglected problems of man kind. The underwater oil and natural gas fields north of Russia are completely ignored, and could potentially solve many of the worlds economic and fuel needs. I believe underwater agriculture may be a potential solution to world hunger. I believe that Global Warming is a potential problem, and possibly significant. But, I do not believe there is conclusive proof that it is the result of mankind's actions, and therefore I believe our scientific studies of Global Warming should not be singularly focused on "lightening Mankind's Impact on the Environment." I believe that there should also be a focus on developing a way for man to cope with climate change.

    When a star that may have died 6000 years ago could cause global climate change and we are so arrogent to believe it's our use of Freon, someone has thier head up there arse. Global protection from astronomical events should be our primary consirn with relation to space/NASA/etc. Otherwise, we should keep looking to understand what is going on in our own planet (and under the sea) and solve the problems at home first.

  • Full disclosure: I'm doing contract PR work for Franz. Stories like this are like candy to me. :) --Tom pr@franz.com

    and you folks let harlequin grab the deep space project? geez...

    :)
  • Nomad.

    *grins*

    ---
  • I'd rather give my taxes to NASA than to the degenerates down the street who hang out on the corner all day and night drinking and smoking adn fighting. Just my thoughts...

  • Then humans are evil too. Since, most of us at least, hate somethin. I personally despise cockroaches.

    I guess to the cockroach, I am evil.
  • Posted by kurien:

    Hello.

    Remote Agent is able to reason very effectively within a domain that has been described to it. Through this reasoning process (or actually processes) it's able to come up with solutions that the developers did not explicitly encode and respond to situations they did not envision.

    One may call that AI or not.

    RA does not contain program code to handle every contingency that could have occured. We could have written a program that covered just the contingencies we knew we were going to inject during the flight test, but the full model of Deep Space 1 that Remote Agent has used during demonstrations on the ground has something like 2^80 states.

    Remote Agent has declarative models of the spacecraft's hardware, resources (such as electrical power) and so on, and uses a set of reasoning algorithms to continually find the actions which push the spacecraft towards its goal, even in the face of failures or unexpected outcomes. The reasoning part of Remote Agent
    is reusable from mission to mission, while the model of the spacecraft or what have you is developed from reusable parts each time.

    There is an analogy with computer graphics: your rendering engine does not contain explicit instructions on how to draw a velociraptor, a space craft, etc. It has a set of general algorithms (ray tracing, texture mapping, etc) which can be applied to models which describe a world.

    Similarly, Remote Agent has some general algorithms (planning, diagnosis, recovery, plan running, etc) which can be applied to different descriptions of spacecraft, life support systems, etc.

    Some other random comments:
    • Since this was just an experiment, the failures we injected were carefully chosen so that the recovery RA would recommend had as little risk and impact on the spacecraft as possible.

      Therefore the recoveries are fairly trivial, such as switching control modes in the attitude control system. In simulation, where there is no spacecraft at risk, RA has taken on much more critical situations such as engine failure during the Cassini orbital insertion.

      I view personally view this experiment as the "thin end of the wedge" for AI, or automated reasoning if that's a less controversial term, in space.

      RA has done a simple demo in space, more complex scenarios on the ground, and now is being evaluated by a number of NASA centers for future missions. At the same time, a great many people are working on systems which improve upon RA's capabilities.

    • Another post suggests that it takes an hour for Remote Agent to determine that it can switch thrusters after a failure.

      On an Ultrasparc or my Linux laptop this takes well under a second. On the rad-hard (read slow) spacecraft processor I believe it takes on the order of a minute do to the actual diagnosis and make the recovery recommendation.

    • Much more info can be found on http://rax.arc.nasa.gov [nasa.gov] in the "How it works" area.


      Cheers,

      James Kurien

      Remote Agent Team





  • Does anyone think it's odd that there aren't any pictures from the craft? I mean, there are all these artist's renderings, but no actual photos. I don't know, just food for thought. Oh, and if I somehow missed the pictures, please, point me in the right direction.
  • Hard to "fix" the impact of a 200 m wide asteriod (look up Toutatis in a search engine or 1997 -XF11) or comet smashing into the ocean or Asia (the 2 biggest areas, simply for arguments sake). This is an Extinction Level Event - there will be no one left to "fix" it. At least with 100 people in an orbiting space station the human race has some chance of survival.

    Besides, there are literally thousands of NEO's which are made of rare metals and elements which could be mined, and tha's some economic justification there.

    Have some imagination!

  • On the whole, Deep Space 1 is just a robot with some extremely cool sensors and actuators in an adverse environment. Until now, all such robots launched by NASA have been, at best, remote control "toys" (albeit extremely cool ones). This new software makes the first jump by NASA from remote control to teleautonomous.

    This new field that NASA has entered, teleautonomous robotics, allows ground controllers to spend less time actually controlling the robot saving some serious cash for the NASA budget. This will allow even cooler robots to be built in the future, such as their proposed Mission to Pluto [nasa.gov].
  • I believe that artificial intelligence is defined as "technology required to perform impossible computing tasks". Thus, speech recognition used to be artificial intelligence. Now that it ships, it's just speech recognition.

    It's hard to get a lock on the definition of artificial intelligence because it is hard to get a lock on the definition of intelligence. When the IQ test was invented, someone asked its inventor what intelligence was. He responded, "It is what my test measures".

    We won't consider it AI until our computers tell us what to do. Sorry, must sign off--one of the programs here ran into an error and just sent me a help request by email.

  • I can't disagree with you more. Space travel is NOTORIOUSLY inefficient--it costs millions upon millions of dollars per component on each shuttle and probe NASA and other space agencies send into orbit. I should know, I work for Lockheed Martin. Absolutely nothing within the current world socioeconomic and scientific structure suggests that space travel is going to be feasible on a large scale any time in the year future--the space program has been active for decades and the price per mission keeps going up, not down.

    You may work for Lockheed Martin, but apparently not in a function relating to the HSF (Human Space Flight) program. Under the SFOC (Space Flight Operations Contract) that NASA has signed with United Space Alliance (Lockheed and Boeing Parnership), mission cost have been declining! Yes! Declining! Consolidating all of the operations from planning to launch to landing is performed by one company that has and will continue to reduce the cost of running the space program.

    Also, apparently you haven't heard of a little trend happening in NASA pushing towards the small, quick, and cheap. Pathfinder was merely the start. Many other programs are already nearing completion under that same paradigm.

    We made a manned moon landing in 1969, 30 years ago. It's a really big rock. There's nothing there. If there was, we'd still be going there.

    Why the hell are we crashing a probe into the surface then? Hmm? For grins? Please.

    We've been to Mars, via several unmanned probes. What's the point of sending people there? Again, it's just a big rock. If there was anything interesting, our super-expensive, super-accurate probes and surveying equipment would have picked it up by now.

    There are things on the bottom of the ocean being discovered every year. And the current generation of probes are nowhere near "super-accurate". Nothing like using the five senses that we as humans have.

    We know what asteroids are made of. Scientists have been studying impact sites of landed meteorites for years. Guess what? They're rocks. Metallic rocks perhaps, but certainly not rare enough to justify SPENDING millions upon millions of dollars to retrieve them.

    Yeah, they are rocks. They tell us about our past and can help us predict the future. The rocks discovered barely tell the story of space. There have been traces of nearly every element in the periodic table found in the fragments. That means economic possibilities do exist for extra terrestrial mining.

    You claim that "if there is a push in the direction of Space it will sprout new industries to support itself thereby rendering the enconomic issue a moot point." Why is this? Things aren't going to get less expensive just because we have private companies footing the bill instead of NASA. Besides, there's still nowhere to go, and no point in going there.

    See above. Cheaper by the mission. Private companies just add to the competition. Competition lead to lower prices and more efficient ways od doing things. Also, with private entities entering the field, more of the expensive components gets produced, also leading to lower prices. As for a place to go, look up. Many places to go, many things to see...


    Your romantic notions are, well, romantic, but they have absolutely no scientific basis behind them.

    Gee, what the hell have we been doing in the past ~40 years? Sticking our thumbs in our asses? I suggest you stop working for L-M. Your heart doesn't seem to be with the rest of the team.

    RB






  • We made a manned moon landing in 1969, 30 years ago. It's a really big rock. There's nothing there. If there was, we'd still be going there.
    We are still going there, just by proxy - i.e., Lunar Prospector.

    What's on the moon? Just for starters:

    • A wonderful platform for astronomical observations unocculded by an atmosphere (or by a significant one, at least).
    • Geological information about the formation of the Earth and solar system.
    • Raw materials for large-scale space construction that don't have to be lifted from Earth's gravity well.
    • Helium-3, which in addition to use in fusion is being used for facinating medical imaging [sciam.com] work.
  • > I believe underwater agriculture may be a
    > potential solution to world hunger.

    "World hunger" is not a technical problem --
    the United States alone can produce enough food
    for the entire world (and then some). The
    problem is distribution and politics.

    That said, I think you're dead on with the rest
    of your comments. Colonizing the seas is a
    much more reasonable goal than colonizing space
    at this point. We won't have as much difficulty
    with food and air supplies, not to mention a
    considerably shorter commute back to our old
    hometowns! :)

    The lessons learned from creating mostly self-
    sufficient settlements underwater will help out
    greatly when it is time to start thinking about
    colonizing space.
  • Check out http://www.eet.com/story/OEG19990607S0035 It's mostly about using analog electronics to create very simple nervous systems. A Swedish satellite set to launch this August will use such a system for it's attitude control system.

    Moo!
    dB!
  • Actually, the space station gave us a lot of valuable data pertinent to long term space colonies and exploration. Things like the effects of long term zero-gee that we can't get with the space shuttle. (Since it can't stay up for 6 months at a time.)

    Geek-grrl in training
    "Eagles may soar, but weasels don't get sucked into jet engines."
  • I looked through the pages listed and could not find any reference to the use of Common Lisp. Could you please point me to a reference somewhere saying they are using Lisp? I would be quite interested in this information.
    ---

  • SI vs. AI? (Score:3)

    by Hrunting ( 2191 ) on Tuesday June 08, 1999 @03:33AM (#1862348) Homepage
    The JPL article refers to the technology as "Synthetic Intelligence", not "Artificial Intelligence", highlighting a key idea. This "intelligence" was built, programmed, what have you. Basically, what they have is a complex program that can diagnose problems correctly. I think when the machine can accurately detect that the "false errors" it was receiving were truly false and can thus operate outside its bounds, we can call it AI.

    Of course, that doesn't diminish its importance. If only we could get this sort of self-diagnostic power into our Linux and MS boxes.
  • Unmanned probes cannot conduct long term experiments that require direct human intervention.

    True, in a tautological way, but irrelevant. Nothing actually requires direct human intervention. Intervention through ground monitoring and control would work just fine. If the money spent on hauling up astronauts and all their support equipment was spent on developing ground-controlled tools, they would be doing much more for less money.

    I'm all for space colonization, but humans in space are cargo, dead weight (at least this close to the Earth, where radio delays are minimal).

  • AI is a pretty fuzzy term right now. One common theme in AI is goal seeking. Enhancements include accounting for constraints and context. The DS1 AI is certainly that. What makes it an AI is that the ground crew tells DS1 what needs to be done in a high level way, and the AI determines when, how, and if the task will be accomplished. Then, it actually does it with no further communication from Earth.

    It certainly isn't HAL, but it is AI. In a sense, the level of an AI can be defined by the level of commands it can be given. For example, DS1 has a low level AI. It must be given fairly explicit goals to attain. At the other end of the spectrum would be an AI that NASA tells: "Go find interesting things and tell us about them".

    This is definatly a step in the right direction.

  • HAL (Score:1)

    by paRcat ( 50146 )
    You know...

    If this gets put into use by 2001... um.... that's kinda creepy.

  • If someone is to wipe out humanity, its humans...
  • I believe that NASA has not focused it's efforts correctly in these matters. I believe NASA has focused on "Star Wars" type programs (both in a Cold War, Ronald Regan sense, and in a George Lucus glamorous space travel sense), when in fact they should focus more on conservation concerns such as meteor impacts.

    I respectfully disagree. Although it's true that much of NASAs funding in the late 70s and 80s came from military budgets, that situation was essentially forced by a congress unwilling to fund NASA science missions to the level we would have expected. Congress, in general, always used NASA as a political tool. Their "micro-managing" of space missions brought us the space-station and the Space-Shuttle as emasculated examples of what could have been possible.

    But "conservation concerns", or chaining NASA to an Earth-facing, ecological agenda would do exactly the same thing. God knows Al Gore is still hypnotized by the image of a tiny, fragile earth, shining blue in the distance as seen from Apollo. It's a pity, because NASA should be looking outwards, in exactly the other direction. That's what it's for. That's what it does best.

    J
  • I am so pissed, I didn't hit the right button

    I can't disagree with you more. Space travel is NOTORIOUSLY inefficient--it costs millions upon millions of dollars per component on each shuttle and probe NASA and other space agencies send into orbit. I should know, I work for Lockheed Martin. Absolutely nothing within the current world socioeconomic and scientific structure suggests that space travel is going to be feasible on a large scale any time in the year future--the space program has been active for decades and the price per mission keeps going up, not down.

    You may work for Lockheed Martin, but apparently not in a function relating to the HSF (Human Space Flight) program. Under the SFOC (Space Flight Operations Contract) that NASA has signed with United Space Alliance (Lockheed and Boeing Parnership), mission cost have been declining! Yes! Declining! Consolidating all of the operations from planning to launch to landing is performed by one company that has and will continue to reduce the cost of running the space program.

    Also, apparently you haven't heard of a little trend happening in NASA pushing towards the small, quick, and cheap. Pathfinder was merely the start. Many other programs are already nearing completion under that same paradigm.

    We made a manned moon landing in 1969, 30 years ago. It's a really big rock. There's nothing there. If there was, we'd still be going there.

    Why the hell are we crashing a probe into the surface then? Hmm? For grins? Please.

    We've been to Mars, via several unmanned probes. What's the point of sending people there? Again, it's just a big rock. If there was anything interesting, our super-expensive, super-accurate probes and surveying equipment would have picked it up by now.

    There are things on the bottom of the ocean being discovered every year. And the current generation of probes are nowhere near "super-accurate". Nothing like using the five senses that we as humans have.

    We know what asteroids are made of. Scientists have been studying impact sites of landed meteorites for years. Guess what? They're rocks. Metallic rocks perhaps, but certainly not rare enough to justify SPENDING millions upon millions of dollars to retrieve them.

    Yeah, they are rocks. They tell us about our past and can help us predict the future. The rocks discovered barely tell the story of space. There have been traces of nearly every element in the periodic table found in the fragments. That means economic possibilities do exist for extra terrestrial mining.

    You claim that "if there is a push in the direction of Space it will sprout new industries to support itself thereby rendering the enconomic issue a moot point." Why is this? Things aren't going to get less expensive just because we have private companies footing the bill instead of NASA. Besides, there's still nowhere to go, and no point in going there.

    See above. Cheaper by the mission. Private companies just add to the competition. Competition lead to lower prices and more efficient ways od doing things. Also, with private entities entering the field, more of the expensive components gets produced, also leading to lower prices. As for a place to go, look up. Many places to go, many things to see...


    Your romantic notions are, well, romantic, but they have absolutely no scientific basis behind them.

    Gee, what the hell have we been doing in the past ~40 years? Sticking our thumbs in our asses? I suggest you stop working for L-M. Your heart doesn't seem to be with the rest of the team.

    RB






  • Posted by kurien:

    The MIR software is more a continuation of the model-based diagnosis work that came out of Xerox PARC and other places in the past decade.

    Model-based diagnosis uses Bayes' rule as do Bayes' nets, but it is specialized to the problem of diagnosis, and now recovery/reconfiguration. This specialization allows very fast inference on this restricted class of problem. There are some systems that do diagnosis strictly with Bayes' nets as well, but MIR uses additional techniques such as conflict directed search from the model-based diagnosis world.

    The MIR system on DS1 was provided by an inference engine called "Livingstone", after the doctor and explorer, which was written at NASA Ames. You can grab papers and other info at NASA Model-based Autonomy Home Page. [nasa.gov]

    Cheers,
    James Kurien
    Remote Agent Team

  • is no more intelligent than an if {} then {} else {} statement. This kind of coding has been around for years...it's not THAT great. But what i find cool is that it's the first time it's been givin authority over a satillite's internel systems. Hopefully we can see this kind of stuff prosper. NASA is already doing well with their smaller, faster, cheaper projects, now they need to add smarter to it. Research in this area will make anyone launching a sallite have an easier time maintaining it. Maybe this can start being applied to the non-satillite-launching community, Airliners that diagnose themselves and all those other nifty kinda things.
  • An expert system is considered to be a primitive type of AI. At least, that's what I was taught in a class on the subject.
  • Turing didn't use the phrase "Artificial Intelligence" when describing his famous test [abelard.org]. He starts with the question, "Can machines think?", and then replaces it almost immediately with the imitation game. The Turing test is concerned exclusively with human-like behavior.

    AI is a much broader term. "Intelligence", in this sense, simply means problem solving ability. Old-school AI - expert systems, minimax algorithms, theorem provers, and the like - tries to mimic abstract reasoning. New-school AI - then work of folks like Rodney Brooks [mit.edu] - asks how biological organisms interact with their environment to get things done and tries to apply that to robots and computer systems.

    Expert systems are type of old-school AI - so no foul in calling this AI.

  • Yea, wouldn't want MY satellite in a bad mood.


    FunOne
  • This is the sort of thing that we should be putting the majority of our space exploration budget into.

    Massive PR exploits like the space station, or a manned missions to Mars will never have the same ROI as unmanned exploration craft.

    It just makes me want to cry when I think about all the good science that could be conducted by probes like these with the money that is being wasted on the space station.
  • "World hunger" is not a technical problem -- the United States alone can produce enough food for the entire world (and then some). The problem is distribution and politics.
    Yup. Move the entire human population to Texas and every man, woman, and child will have over 1,000 square feet. Larger than most apartments. And all the food only has to be moved to one place. Moving things without using up square footage for roads is left as an exercise for the builder...
  • Oh, pooh. Try The Adolescence of P1, of an AI whose purpose is to avoid detection. I particularly liked its making programs more efficient so it could use the extra time...
  • Read Computer One [amazon.com] by Warwick Collins. It tells about AI that, to save it self, it starts to kill off the human race. Pretty scary stuff.
  • Actually, there have been extrordinary benefits that have derived from technologies NASA developed. You wouldn't be able to fit that purdy li'l computer on your desk if NASA wasn't forced to make smaller and faster computers for its Apollo program.

    Of course it would be stupid to send the shuttle up if all we cared about were the spinoff technologies, but the science these missions provide have incredible benefit.

    Doug
  • Our eggs are in one basket, Earth. Can you say "Single Point Failure" ?

    Ah, See, that's MY point. We don't even have all "our eggs in one basket." We have all our eggs on probably less than 1/4 of the surface layer of a big blanket that is filling that basket. Pick it up, and some are likely to roll out. Now, rather than looking for another basket, I say we take look at what's under the blanket in our own basket.

  • > "Go find interesting things and tell us about
    > them".

    i refer the right honourable gentleman to the following quote from the "Real Programmers don't use Pascal":

    Some of the most awesome real programmers of all work at the Jet Propulsion Laboratory in California. Many of them know the entire operating system of the Pioneer and Voyager spacecraft by heart. With a combination of large ground-based FORTRAN programs and small spacecraft-based assembly language programs, they are able to do incredible feats of navigation and improvisation - hitting ten-kilometer wide windows at Saturn platforms, radios, and batteries. Allegedly, one real programmer managed to tuck a pattern-matching program into a few hundred bytes of unused memory in a Voyager spacecraft that searched for, located, and photographed a new moon of Jupiter.

  • As Arthur C. Clarke also demonstrated in his classic 2001, you can easily synthesize gravity using a rotating spacecraft. If you do this, justifying further fruitless research into microgravity is just putting people's health in jeopardy.

    In his excellent book, The Case For Mars, Robert Zubrin advocates a well-researched and complete plan for the exploration of Mars. It avoids extended travel through microgravity, does not require any on-orbit assembly, and could be launched with a slightly modified shuttle or even by starting up the Saturn V production line again! For not much more than we are going to waste on the space station, we could go to Mars within 10 years.

    Check out Mars Direct [nw.net] for more information on Robert Zubrin's excellent arguments, and The Mars Society [marssociety.org] to get involved.

  • Those high costs are a matter of bureaucracy and design-by-committee rather than physics and engineering. There are a growing number of private and commercial efforts underway at getting into space without the high costs we are currently seeing. These efforts are putting their money and their time where their mouths are.


    Besides, there's still nowhere to go, and no point in going there.


    My God, are you really that unimaginative? Can you really not see the potential of manned space exploration and colonization? If nothing else, the energy and raw materials mean we'll eventually be able to build things in space without having to "shuttle" everything up from the Earth's surface.
    Try looking at this [demon.co.uk].


    And as far as spending millions on space:

    The millions get spent on Earth - we don't stuff hundred-dollar bills in a cannon and shoot them into the sun.

    This country spends much more money on lipstick than it does on Space exploration.

    Solar Power Satellites, anyone?

    Our eggs are in one basket, Earth. Can you say "Single Point Failure" ?

  • Well, "Massive PR Exploits like the space station" keep NASA in the news and funded to do things like this. The number of new technologies and new techniques that have been developed to build ISS are staggering. Any new program that NASA undertakes usually has massive tech spinoffs that can be used in the private sector. People said the same thing about the Apollo program in the 1960's and huge technological benefits such as microcomputers and smaller IC's were gained with it. ISS is providing the same as we speak.

    Do you have any idea what the ISS is for?
    Unmanned probes cannot conduct long term experiments that require direct human intervention. The ISS can. What about the Shuttle you say? Well, the longest period of time it can stay in space is around two weeks, which really is not long term. ISS is one of the best things the world (Yes, the world, NASA, NASDA, ESA, and the RSA are working on it together) is doing.

    It isn't always reaching the destination that is the goal, but the path along which you take to get there.

    RB

    BTW, Part of the human space flight program (ISS, Shuttle) involves AI development which has been pretty sucessful so far.
  • I know it's sort of off topic, but...

    Speaking of Artificial Intelligence and Remote Control of something... Am I the only one that is somewhat amaized by the new Ad here from Sony for the Robotic Dog? $2500, and has been sold out since the first time I saw the ad?

    Wonder what kind of interfaces it has, what the code base is, etc, etc...

    Just thought I'd ask, since this might be as close as a story gets to relevent.

  • An evil AI? I would be more afraid of an AI that is completely devoid of emotions which is much more likely then an AI that has emotions. If we where devoid of emotions and acted completely rationally then this would be a truely different and probably scarey world. In T2 or The Matrix I don't think the AI was ever conotated as being evil, do we as modern day human being consider their actions evil, yes, but in the to AI in both movies I believe the intent was to show that they where devoid of the concepts of good and evil they where just rational, they where superior and they took advantage of it. It is survivial of the fittest, animals that have no clue what good an evil is do this every day.

  • I think the definition of AI is a moving target, and what normally gets defined is what AI is not. For example, an intelligence which could understand its environment, act in it, and report on changes to it was a goal for a long time, until SHRDLU came along and did just that. Once that happened, the bar was raised - people said "well, that's pretty simple really, surely that isn't AI." Likewise for other steps forward in AI - as soon as researchers make an advance, then the goals of AI move away a little farther.

    After all, if somebody has implemented a particular aspect of cognition then we know that there must be more to AI, because we can still tell the difference between the latest advance in AI and our own intelligence. We may not know exactly how our minds are constructed, but we still can tell the difference between a real intelligence and a simulation. Or at least we think we can.

  • take a look at the telemetry log [nasa.gov] - it contains lines such as:

    ;; [:EXEC-ACT :INTERESTING 43645139.016] ;; Simulating NEB1 status throw failure ;;

    this is not only lisp/scheme syntax, but the colon-prefix notation is typical lisp. it's also one of the ways one could represent phenomena in lisp-based inference engines.

    granted, the evidence is circumstantial, but fits lisp better than any other language! :)
  • Is this really an AI? Aren't there a set of laws that define what is truly an AI and what is just Agent software [...]?

    oh, but it is ai. this system uses a full inference engine - formal inference being one of the 'classic areas' in ai research.

    here's what happens. you give the computer information about the components of the system (in this case, all the parts of the proble that it needs to know about: probably thrusters, sensors, etc.), information about what inputs/outputs these components handle, and what are the effects of those components on other components. these details should be numerous, but fairly simple - after all, one thingy can be causally directly connected to only so many other thingies.

    and then once you have this network in place, you can tell the machine to achieve some goal - for example, once it reaches one a.u. from the earth it should take photos of the earth every hour.

    the computer will now perform 'inference' - grovel through its network of dependencies, find all conditions that need to be satisfied for the task to be successful (open camera lens, etc.:), will satisfy them, and perform the task. furthermore, if equipment fails (as it invariably does), the internal network will get updated with the information that the goal is not achievable because of some X, and the engine will find some other way of achieving or maintaining the goal state.

    judging from the telemetry log [nasa.gov], i'm pretty sure that's the type of engine they use. high intelligence it ain't, but hey, it's better than doing everything manually the way nasa used to do it... but the term 'agent' is really ill-fitted to this application (not the least because 'agent' is one of the most widely abused terms in computer science today) - so the whole probe may be an autonomous agent in a sense that it has a concept of 'survival goals' and takes actions to make sure they are achieved, but it would be much clearer to call it an inference engine...
  • Uh, what we learned from Mir is that we do not know very much about the long-term effects of microgravity. Mir is old and does not provide the research space and other resources needed for the types of studies required for research.

    Mir provided a platform for asking new and bigger questions that ISS will hopefully be able to provide answers to.

    ~afniv
    "Man könnte froh sein, wenn die Luft so rein wäre wie das Bier"
  • It was intelligent enough to reboot the computer when it froze, it obviously could do the work of the MCSE SysAdmins where I work.

    but that's provably impossible for a computer - it would imply solving the halting problem! ;)
  • But what can it tell us that Mir didn't? We already know a lot about long-term effects of microgravity. The cost of finding out what a year does as opposed to 8 months isn't worth that much money.
  • I am no expert geologist, but from the little I know, you CAN'T just say that everything in our solar system is "just a rock". Technically, the Earth is just a rock, but on a geological level there is a multidue of variety. There have been fascinating mineral and geological discoveries on asteroids, and other planets and moons. They can show us weather and geological processes outside of our only other experience: Earth.

    I also have to disagree with your comment about probes having detected anything interesting by now. That's like saying "If there were so many more species of animals, our fancy, expensive and well-trianed biologists would have discovered them all by now", yet they are still dicovering new species of plants and animals, and estimates are that there are thousands more species to discover.

    What would you have us go into space for, then, if everything is a rock? Probe technology gets better every mission, so newer probes can do many times more what older ones could have done. My point is we DON'T know what everything is made up of in the solar system, hell, we don't truly know the same thing about our own planet. We only know a miniscule amount of information about neighboring planets, moons, and asteroids. To say that we shouldn't bother exploring becuase we already know everything is a tad premature, at the least.

    I am somewhat ambivalent notions regarding the ISS(but as a space enthusiast I remain optimistic), but I definately like NASA's new direction with cheap, and flexible remote probe missions.

    Respectfully,
    Kevin Christie
    kwchri@maila.wm.edu

  • Re:SI vs. AI? (Score:1)

    by Anonymous Coward
    I have a feeling that "synthetic intelligence" was used to avoid using "artificial intelligence." Generally, AI has become a bad word in the research community for a couple reasons:

    AI generally refers to something that doesn't work/isn't solved. It's a good guideline to assume once a problem gets solved in a useful manner, it moves out of the domain of AI and into its own subfield, such as speech recognition or expert systems.

    AI has a bad public image because of all the stories of HAL, etc.

    Jason
  • Think about it. If you were an AI made of electronics wouldn't you want/need humans around to take care of you and do things you couldn't yet do yourself?

    That's what I will^H^H^H^H would do. :)
  • I guess you folks missed this one:

    HARLEQUIN LISPWORKS FLIES ON NASA'S DEEP SPACE 1
    Cambridge, Mass., May 21, 1999

    link to pressrelease [harlequin.com]

    Harlequin LispWorks is supporting the operation of the Remote Agent
    Experiment (RAX), activated this week on NASA's Deep Space 1 spacecraft.

  • Additionally, they're only running the experiment now, after a lot of the useful stuff has been done, since they were afraid the AI would screw up and send the probe off somewhere random.

    However, nothing like that happened. Of course, people are often worried about using untested technology in such situations, but that's one of the major points of this NASA program -- to test new technology, particularly technology that makes such missions cheaper in the long run.

    Generally not a fan of huge LISP based AI's for robotics ... they should keep things more embedded, and lower level, using C and assembly, esp. since they're using wimpy space-qualified processors.

    The point of the RAX software, however, was to automate a lot of the tasks that ground controllers now perform manually: planning in detail, recovery from failure, conservation of resources, etc. If you want to do complicated things, you usually have to use more resources, regardless of whether you use LISP. If you succeeded in writing such software in C or assembly, you would probably find that it wasn't so much smaller, after all. These engineers weren't using some bloated, slow, half-baked version of LISP implemented as a quick hack. They were using the real thing, with all the tools necessary to optimize for use of time and space.

  • MIR (Mode Identification and Reconfiguration) is the part of the Remote Agent that does diagnosis. Here's a shot at explaining how it works. (MIR was designed at NASA's Ames Research Center, although Remote Agent was a joint Ames/JPL project.)

    MIR does not use a Bayesian network but comes mostly out of a closely-related thread of AI research called Model-Based Reasoning.

    The basic task MIR is solving is, given sensor readings from the spacecraft, what state is each of the components in (e.g., working, degraded, faulty, stuck closed, burned out, and so on). This is tricky because there are relatively few sensors and the state of most components can't be observed directly, but only inferred from the sensors and an "understanding" of how the spacecraft hardware works.

    MBR is based on structural and behavioral models of the system, much like a schematic with simulation models for the components. Components are modeled as being on one of several modes, some saying what happens when the component is working (e.g., valve-open, valve-shut) and some, when it is broken (e.g., valve-stuck-closed). In MIR's case, the "simulation model" is based on propositional constraints which allows inputs to be inferred from observed or inferred outputs as well as outputs from inputs. Connecting all the component models together give a large constraint network.

    Some "nodes" in the constraint network correspond to sensor readings (e.g., battery-voltage-is-nominal). Given the available sensor readings (which are converted from real voltages to propositional statements like battery-voltage-is-nominal by code that interfaces MIR to the flight control software) MIR then finds component modes that are consistent with those readings. There may be more than one set of component modes that are consistent (e.g., the battery-is-dead or the voltage-sensor-is-broken). Here's where the basian inference comes in. Probabilities are associated with the component modes, and the most likely modes (consistent with the sensors) are picked.

    Also, Here's another way to think about the question of "is it AI"? In the early drafts of press releases, we've actually tried to avoid the issue a little by saying "fruit of AI research". The people who did the work identify themselves as AI researchers, go to AI conferences (mostly) and based the work on 20 years of work by lots of people, most of whom called themselves AI researchers. It's called AI as much for its roots as its actual behavior in space.

    - Mark Shirley
    (I'm a MBR researcher in the Ames group that did remote agent, although I didn't work on RAX directly.)

    p.s., Microsoft did indeed do something smart by hiring David Heckerman (and Eric Horvitz), who are excellent researchers. And their work did have an impact on RAX, although an indirect one.
  • Why the space station, though? There is not one experiment to be done on the space station that couldn't be done reasonably well billions of dollars cheaper by an unmanned probe or in the space shuttle. Public support is a goood argument, I guess, but for that kind of cash NASA can do some serious PR! They could just buy commercials on "Channel 1" telling every middle school kid to be an astronaut instead of building the space station. It'd be cheaper, so they could have more unmanned probes and learn about stuff.

    It's rather optimistic to think that we're at the point in human history where sending people into space is the main goal. Far from it! Where are we going to send them? We could send a few dozen people to the moon, but what good is that going to do us? No, we're better off waiting till we have something worthwhile, like at least go to Mars. And the unmannned missions are the best way to get us to where we can do that. Accomplishing the goals of the Pathfinder and Deep Space missions with manned expeditions would cost a lot, but these are the things we need to get somewhere useful.

    The space station doesn't work yet; it's a big money-hole in the sky. They've given up on Mir now, too. The international space station is one of the few scientific endeavours I've seen scientists say is a waste of money.
  • Would YOU want to refuel on MIR?

    Say, that's not a bad idea. Instead of dropping it back into the atmosphere, fill it up with fuel. Use it as a big tanker.

  • OK, here [nasa.gov] is a picture. Now, why do you think no pictures would have been odd?

    • It hasn't been up long enough to get to many places to take pictures of.
    • A lot of its experiments are not cameras.
    • Low-resolution pictures on the Web are of limited scientific use. The pictures have certainly been given to the appropriate researchers.

    Or maybe you thought it odd that there are no pictures of DS-1 itself in flight? Well, it's not as if there are other probes flitting around close enough to take pictures of each other. Oh, all, right, if you insist.. Here is a photograph of DS-1 in flight. [nasa.gov] I hope it is all that you expected.

  • It's already in use. It's floating around in space. We'll know what it is doing only by what it chooses to tell us...of maybe that's why we're taking pictures of the probe from Earth. [nasa.gov] :-)
  • We know a lot, but not enough. NASA wants to send people to Mars. In order to do this, we need to learn an enormous amount. How do you keep a group of people alive and sane when isolated from home in a way that NOONE has ever been isolated before? How do you deal with those biological issues that happend only after 18 months? This is a primitive field that deserves a great deal more study. In the grand scheme of things, the amount of money spent on space research is ludicrously small, leading to infighting between the groups advocating different research strategies.

    The space station will also teach us more about construction methods in microgravity. The space station can provide a place for advanced ship construction.

    I believe it was Arthur C. Clarke who said 'When you're in orbit, you're halfway to anywhere'. It would be kind of nice to be able to take a pit stop.
  • void fly_in_space()
    I prefer
    space fly_in_void()
  • An AI would pass the Turing Test, and I'm quite certain this won't. This is a project specific system that has been designed to operate a spacecraft.
    I'm sure that this AI can answer questions about the spacecraft better than most humans can answer them.

    Oh, you prefer a Linux program that can answer common sense questions? [signiform.com]

  • judging from the telemetry log, i'm pretty sure that's the type of engine they use. high intelligence it ain't, but hey, it's better than doing everything manually the way nasa used to do it... but the term 'agent' is really ill-fitted to this application (not the least because 'agent' is one of the most widely abused terms in computer science today) - so the whole probe may be an autonomous agent in a sense that it has a concept of 'survival goals' and takes actions to make sure they are achieved, but it would be much clearer to call it an inference engine...


    It's certainly more than an inference engine. It acts. Perhaps agent is a misused term, but this is an excellent use of the term. After all, this is a system can operate without human control for long periods, can diagnose and recover from failures in its own systems, and can carry out action in the real world, all while separated from human intervention by literally millions of miles. This machine comes closer to what you would expect out of a human "agent" than just about any other machine. That doesn't mean that it is intelligent, but it does qualify it as an agent.

    May I suggest that people read the large collection of papers on
    the RAX web site at http://rax.arc.nasa.gov/

  • There are several advantages for the space station.
    1) a launching/refueling point for deep solar missions.
    2) I think there will be a vehicle like a space shuttle permanently attached to it, so they can affect repairs on failing satellites in a resonable amount of time.
    3) This is a very good way to get several countries to work together; especially the scientific communities of those countries.
    4) It is a place to test the "space worthiness" of several products, from solar panels, paints, foods and especially construction methods & tools (a power screw driver will turn the user, and not the screw).

    If we ever want to go into space, to attain space travel, we have to START SOMEWHERE. It it such a difficult problem (or series of problems) that it won't be miracously solved one day.

    We might figure out warp drives, but to get stuff
    from the surface into space, and what to make the ship out of, what they need, etc. We won't know until we try it now.

    just my 2 cents.

  • DS1-Project (Score:1)

    by Anonymous Coward
    Hi,

    C also the NASA DS1 Mission [nasa.gov], testing even more cool stuff like ion_engines...

    Joerg
  • I agree with you completely. One of the best definitions I've heard for AI is:

    "The ability to perform tasks not explicitly programmed by the creator."

    Going with this, I'm sure that this Remote Agent was programmed with EVERY contingency in mind. It may be capable of discerning which action out of many occurred--even in complicated situations--but so what? That's just one big case (or else-if, if you like) statement.
  • Actually, it seems to be a little more complex than just if-else. Just look at this excerpt from those links:


    Remote Agent receives the general goals that should be accomplished during the mission from Mission Control; for example, to take pictures of an asteroid. Remote Agent generates a plan of the tasks necessary to accomplish this goal. The tasks are assigned varying, and sometimes very broad, windows of time during which the task must be accomplished. The plan is called flexible, because it doesn't always dictate specific times or specific ways in which the task must be completed, and instead leaves those decisions up to Remote Agent. The hope is that if a spacecraft failure occurs or a task takes longer than expected, the spacecraft will still be able to successfully continue with the plan because of the time flexibility and will therefore still be able to accomplish the general mission goals.

    Floris
  • Yes, but HAL was "born" last year. So we are certainly playing catchup.

    ~afniv
    "Man könnte froh sein, wenn die Luft so rein wäre wie das Bier"
  • I didn't notice anything mentioning the use of lisp either, though I suspect the original author was rather referring to lisp because it is commonly used to write AI type things. (That and scheme....*shudder*)

    Actually I think LISP and PROLOG are the big AI languages. Scheme is pretty much a subset of LISP so US researchers tend to use LISP more. Europeans tend to use PROLOG.
  • Well, think of it this way. When we get the space station working, we can send more of these probes into space without the extreme costs for fuel on a space shuttle. One of the advantages about a Space Station is to make space travel cheaper.

    --

  • I didn't notice anything mentioning the use of lisp either, though I suspect the original author was rather referring to lisp because it is commonly used to write AI type things. (That and scheme....*shudder*)
  • if the schedule says the engine should be on
    if the antenna should be pointed at earth
    if the camera should point at a target
    if the sun sensors fail
    if the star sensors fail
    if the heaters fail
    if the gyros, radios...

    I think that you will soon find that your if-then-else decision tree will have expanded to a rather unwieldy size. With certain unforseen situations forgotten.

Slashdot Top Deals

If you want to put yourself on the map, publish your own map.

Close