Slashdot Mirror

← Back to Stories (view on slashdot.org)

Satellite Loaded With AI For Self-Diagnosis

Posted by timothy on from the would-it-prevent-ben-afleck-movies dept.

TheReckoning writes "NASA has loaded its E0-1 Satellite with Artificial Intelligence to diagnose on-board failures. The software 'works by comparing a computerized model of how the spacecraft's systems and software should perform against actual performance. If the spacecraft's behavior differs from the model, then the ... "reasoner" looks for the root cause of this difference and gives flight controllers several suggestions of what might have gone wrong.' Another NASA probe loaded with AI was Deep Space 1."

7 of 218 comments (clear)

  1. So this may be a simple question but... by EvilSS · · Score: 4, Interesting

    ..in all seriousness, what happens if the AI system malfunctions?

    --
    I browse on +1 so AC's need not respond, I won't see it.
    1. Re:So this may be a simple question but... by Brett+Buck · · Score: 5, Interesting

      Bingo - you have hit that nail on the head. These sorts of systems in aerospace applications are absolutely notorious for detecting proper (but off-nominal) operation as a failure, and then going off and reconfiguring a bunch of stuff unnecessarily. Or diagnosing real problems incorrectly, and either not helping or making things worse, or much worse.

      Even more importantly, the testing associated with these systems is very expensive and time-consuming - which means they don't really test it very well at all.

      I've seen similar systems in action in real space flights - and for the most part, it just makes things worse. If you were to limit yourself to simple things you really could detect, it would work out fine for the most part. But the tendency is to make it try to be a magic fixit device for any problem that comes up.

      In one case, I saw such a system deploy an appendage in conditions that resulted in the spacecraft structure being severely damaged. In another, it reconfigured every spacecraft system to the redundant unit in response to a trivial problem - when all that would have been required would have been to wait 20 minutes, then correct the trivial problem.

      Brett

    2. Re: So this may be a simple question but... by shadow_slicer · · Score: 2, Interesting

      There are conditions under which that sequence would be provably finite. If you assume that the AI necessary to monitor another AI is simpler, then eventually you will get to an AI that is infinitely simple.

  2. Wouldn't obvious failures be detectable anyway? by mind21_98 · · Score: 4, Interesting

    Wouldn't obvious failures (like the failure of a sensor) be detected by Mission Control without LV2? Or is LV2 more along the lines of a troubleshooter application for your computer, where you specify the problem and it gives you advice?

    --
    US businesses that currently accept chip and PIN/signature
    1. Re:Wouldn't obvious failures be detectable anyway? by EmbeddedJanitor · · Score: 2, Interesting
      There is much sense to what you say here. After all, it is easier to write/ modify/ debug/model a bunch of things on planet earth than it is to do this with a little itty-bitty piece of software running in a probe.

      However, to do such modelling etc on the ground typically means pumping a hell of a lot of diagnostic trace stuff to earth for analysis. Likely more than can be accomodated on the link. For this reason, some AI stuff could help to identify the problems faster and allow further in-depth debugging of what is wrong.

      The skeptical view is also that the AI group has got to fly something to justify their budget.

      --
      Engineering is the art of compromise.
  3. We're talking about two different things. by MOMOCROME · · Score: 2, Interesting

    What you are trying to describe is some sort of conciousness, or more particularly sapience, which is our (human) flavor of conciousness. There are many other types of intelligence, artificial or not. Even those books you mention take pains to describe this, especially in the description of the ant colony problem from GEB.

    It would be a simple thing to crash an ant colony with recursion using only an eye-dropper full of the right pheremone(s). Does this mean the ant colony has no intelligence? Not at all. For another example, consider schizo-effective disorders and autism: these are obvious malfunctions with the recursion control mechanisms in the human brain. But I'd scarcely describe the poor victims of these awful conditions as lacking 'intelligence', 'conciousness' or 'sapience'. They just have a bug in the code, and your haughty dismissal of intelligent systems for lacking this capability smacks of some pretty cruel callousness.

    The same principles are at play with these early examples of machine intelligence. That is to say, they can be intelligent without achieving conciousness, or can achieve conciousness while being in peril of recursion loops.

    But the overarching point here is that putting some of the fundamental building blocks of machine conciousness into service, like they are in this control system, is a substantial step in the drive to get to where 'laymen' like yourself can finally be impressed.

  4. I wonder if my code is in space now. by naoursla · · Score: 3, Interesting

    I worked on a project similar to this for NASA's interferometry telescopes at JPL a summer ago.