Slashdot Mirror

← Back to Stories (view on slashdot.org)

Curiosity Rover Decides, By Itself, What To Investigate On Mars (sciencemag.org)

Posted by BeauHD on from the autonomous-robots dept.

sciencehabit writes: NASA's Curiosity rover landed on Mars in 2012, in part to analyze rocks to see whether the Red Planet was ever habitable (or inhabited). But now the robot has gone off script, picking out its own targets for analysis -- precisely as planned. Last year, NASA scientists uploaded a piece of software called Autonomous Exploration for Gathering Increased Science (AEGIS) adapted from the older Opportunity rover. Curiosity can now scan each new location and use artificial intelligence to find promising targets for its ChemCam. Compared with the estimated 24% success rate of random aiming at picking out outcrops -- a prime target for investigation -- the current version of AEGIS lets the rover find them 94% of the time, researchers report.

10 of 73 comments (clear)

  1. Rover, FETCH! by war4peace · · Score: 2

    "We're too lazy to tell it what to look for, so we just let it do its thing and present that as an achievement".
    But seriously, interesting story.

    --
    ...gis sdrawkcab (usually not responding to ACs; don't bother posting as AC)
  2. 94% of what? by Mosquito+Bites · · Score: 5, Insightful

    Headline claimed that the 'success rate' is 94% Went to read the summary to find out what the 94% is all about but unfortunately it isn't saying much Any clue? SVP?

    1. Re: 94% of what? by arth1 · · Score: 2

      If they put aman on Mars I dont care. Its a waste of money.

      Well, yes. Rice is not a prime candidate for cultivation on Mars, even hardy winter crops like aman.

  3. Number Five Is Alive! by Solandri · · Score: 4, Interesting

    Original concept art for Curiosity

    Number Five from Short Circuit

    Things that make you go hmm...

  4. Twaddle by Hognoxious · · Score: 5, Insightful

    gone off script, picking out its own targets for analysis -- precisely as planned

    If that was planned, then by definition it's not off script. If a music score says "imrovise" or a sc-fi script says "// technobabble here" then that's what was planned.

    --
    Confucius say, "Find worm in apple - bad. Find half a worm - worse."
  5. Re:Crash in a parked trailer by KiloByte · · Score: 3, Funny

    Will it autonomously crash in a parked trailer ?

    That's why you run such a prototype 390 million km from the nearest trailer.

    --
    The creatures outside looked from Alt-Right to Antifa; but already it was impossible to say which was which.
  6. Remember this is "weak AI" by gweihir · · Score: 2

    I.e. the "AI" with no "I" in it. Before the demented AI hype, this was called "automation".

    --
    Most ACs are not even worth the keystrokes to insult them. Be generically insulted by this and ignored otherwise.
    1. Re:Remember this is "weak AI" by angel'o'sphere · · Score: 2

      Automatic route finding *and* following it, especially with several propelled wheels in an difficult environment (sand / slippery ground etc.) was always considered weak AI, or part of the 'field of AI'.
      I worked ariund 1989/1991 in robotics with self driving robots. They had several 68040 CPUs and ultra sonic sensors to measure their position. Four wheels that had diagonal rolls in their tires, so by rolling the front wheels backward and the rear wheels forward the robot would shift because of the rolls to one side (without turning itself and without turning the wheels). letting the wheels rotate into the other direction the robot would shift to the other side.
      All that was supposed to happen in a controlled working area for the robot. So actually they could have given him a previse X/Y coordinate about its position from the outside. But they wanted all at once: perception, building a model of the environment, planning routes through it, following the route and checking its position continiously on that route.
      Anyway: the bot was suoer maneuverable (I think the wheels are called Mecano or Mercano or something), and super slow. It basically stopped every 30cm and took over a second to measure its position again. Not sure how much he had to (re-)plan afterwards again. Point is: it basically never was at the position its internal modle and its driving actions predicted.
      This all might not sound like AI to the AI bashers here on /.
      However endless PhD's and some Professors were produced during that research project (Professor Rembold, meanwhile deceased, at KIT Karlsruhe was one of the heads of that research).
      It was hard work. A joint venture between several industries, and institutes from several univeristies working in the are of 'C ognitive Systems', 'Planning', 'Automation' etc. and at that time people considered it most certainly AI.

      Today we stand on the shoulders of those giants and others .... now we consider it simple, OTOH that robot had only 1 or 2 MB of RAM. I believe 2 68040 CPUs in the 32MHz range.

      My iPhone 4 probably has 10,000 times the computing power of that robot (to lazy to calculate it exactly).

      --
      Cost free eBook I read (by iBook/Kobo/Amazon/ObookO/Gutenberg etc.): "The Green Odyssey" by Philip Jose Farmer.
  7. Re:What could possibly go wrong? by stooo · · Score: 3, Funny

    Oblig XKCD :
    https://xkcd.com/1504/

    --
    aaaaaaa
  8. Re:What could possibly go wrong? by rgbatduke · · Score: 2

    The question is, can it use its laser to smelt metals and create spare parts, additions, and (one day) rockets? Will it join up with other Mars rovers we might send and recombine their onboard plans and manuals to build new little roverlets equally well equipped?

    If so, one day I hope to greet our new Mars Rover overlords!

    --
    Even when the experts all agree, they may well be mistaken. --- Bertrand Russell.