Slashdot Mirror


Student Makes Real-Life Portal Turret

lukehopewell1 writes "A Penn State robotics student has gone to the effort of building a working, automated turret from the video game series Portal. Powered by a webcam, Arduino boards and hacked up USB-missile launchers, this is one serious piece of kit that is just as adorable as its in-game counterpart. "

24 of 91 comments (clear)

  1. Not enough bullets by ebunga · · Score: 4, Funny

    And it doesn't even fire the entire bullet.

  2. Yes by synapse7 · · Score: 5, Insightful

    All military gun turrets should have a seductive female voice enticing their targets.

    1. Re:Yes by HeLLFiRe1151 · · Score: 4, Funny

      It's from Penn State, what else would you expect?

      --
      I've got 101 mod points and you can't have them!
  3. Awesome! by scubamage · · Score: 4, Interesting

    Great job kid! Keep it up and you could have an awesome job at the army corps of engineers, or at some other defense contractor. Great work.

    1. Re:Awesome! by mblase · · Score: 2

      Great job kid! Keep it up and you could have an awesome job at the army corps of engineers, or at some other defense contractor. Great work.

      Um... did you know that the Army Corps of Engineers has almost nothing to do with defense contracts?

  4. Whyyyyy? by GameboyRMH · · Score: 3, Insightful

    I don't blame you.

    --
    "When information is power, privacy is freedom" - Jah-Wren Ryel
    1. Re:Whyyyyy? by buchner.johannes · · Score: 2

      and why is the robot shooting for your balls? I think it's trying to stop you from breeding ;)

      --
      NB: The message above might reflect my opinion right now, but not necessarily tomorrow or next year.
  5. Mini-gun by Sperbels · · Score: 4, Interesting

    I think a rubber band mini-gun would have been more suitable for this project. http://www.youtube.com/watch?v=v7UIoYGSVoQ

    1. Re:Mini-gun by GameboyRMH · · Score: 2

      I was wondering if he would replicate the firing mechanism shown in the game, where bullets are loaded from a hopper and shot out using an electric firing mechanism (no powder required). Now that would be impressive.

      --
      "When information is power, privacy is freedom" - Jah-Wren Ryel
  6. $5 says by future+assassin · · Score: 2

    There's a GPS devices underneath his car now and $10 says his next flight he's gonna receive a complementary ass exam.

    --
    by TheSpoom (715771) Uncaring Linux user here. I have nothing to add to this but please continue. *munches popcorn*
  7. And? by Anonymous Coward · · Score: 4, Insightful

    Is this of special note because its slightly reminiscent of Portal? These guys put together an amazing tracking turret and stuffed it with paintballs:

    http://www.youtube.com/watch?v=RxBa5bQfTGc&feature=player_embedded#t=62s

    http://www.paintballsentry.com/

    Its tracking and rate of fire is pretty terrifying, in fact. I don't see why it couldn't be sleeved to look like a Portal turret too and be way more impressive.

  8. Needs a lot of work. by Lumpy · · Score: 4, Interesting

    Tracking is really slow. it should be tracking so fast that you dont get the jerky motion.
    Add in predictive aiming and you have something that may get you a job for life working at a military contractor.

    --
    Do not look at laser with remaining good eye.
    1. Re:Needs a lot of work. by Stormthirst · · Score: 2

      Provided you can persuade the to wear pink shirts.

    2. Re:Needs a lot of work. by Anonymous Coward · · Score: 2, Informative

      I go to a Penn State branch campus. Considering that the main campus only gives each person $1000 and one semester to design and build their project, you cant say that this is not incredible! Yes it can certainly be better, but with the monetary and time restraints you can not argue with the final result!

    3. Re:Needs a lot of work. by tibit · · Score: 2

      That's what you get if you slap together a proof of concept. Matlab's image processing can run much faster than demonstrated here, even on modest hardware. It's an unfinished hack, yes.

      --
      A successful API design takes a mixture of software design and pedagogy.
    4. Re:Needs a lot of work. by ArhcAngel · · Score: 2

      it can certainly be better, but with the monetary and time restraints you can not argue with the final result!

      This is /. We can argue about anything...like the use of contractions or garamond is superior to times new roman.

      Watch...I'll get numerous replies telling me I'm an idiot for even suggesting garamond is a usable font.

      --
      "A person is smart. People are dumb, panicky dangerous animals and you know it." - K
    5. Re:Needs a lot of work. by msobkow · · Score: 2

      I agree. Way back in 1988, one of my university cohorts developed image analysis and recognition software that would identify a target and paint a crosshair mid-torso. Of course in the age of the VAX 780, it was far from real-time processing. But given that even "slow" computers are over 1000 times as fast nowadays, there is absolutely no excuse for the apparent processing delays in this project. (It used to take an hour to do the analysis and paint the crosshair on a graphics terminal, so with 1000 times the processing power it should be able to do it at least 16 times a second on even a P4.)

      Maybe they made some bad choices for their software. My friend's work had been raw C code, not a higher level package or interpreter like MatLab.

      --
      I do not fail; I succeed at finding out what does not work.
    6. Re:Needs a lot of work. by msobkow · · Score: 2

      The '88 software didn't require any cheating like using a bright pink shirt so it would stand out from the background, either. It would track anything human-shaped that crossed it's path, and was "smart" enough to reject other 4-limbed creatures like cats or dogs.

      Rather than centering on "the pink blob", it used some interesting algorithms to build up a stick figure of the object, identify the torso stick, and aim for the middle of the torso stick.

      They had a lot of fun on that project. Pretty much everything except the software used to drive the graphics terminal and do the image capture had to be custom coded.

      --
      I do not fail; I succeed at finding out what does not work.
  9. Re:Really slashdot? by wjhoffman1983 · · Score: 3, Insightful

    I'm an idiot and I couldn't mock this up.

    It's obviously not professional/military grade but I still found it impressive. Are you the type to mock amateur artwork because it's not Monet or Rembrandt?

  10. Better example by pshuke · · Score: 2

    While definitely cool project, I think there are better examples out there. http://www.youtube.com/watch?v=6QcfZGDvHU8#t=16 also shows an auto tracking turret with portal voices. However in this case the tracking seems quick enough to be useful, and it actually hurts people. Add some white plastic, and you're pretty close to the "real" deal.

  11. Re:Really slashdot? by genghisjahn · · Score: 2

    If we quit putting cool proof of concepts on slashdot, there wouldn't be much content besides apple/google/microsoft bashing. And he didn't post a critique. It's just name calling.

    --
    Sorry about the mess.
  12. Needs a Nerf Vulcan and neural networks by SiferDomm · · Score: 2

    With artificial neural networks, stereoscopic vision, and a Nerf Vulcan you can:

    Shoot down an RC helicopter
    Track and range a helicopter
    And even compensate for bullet drop

  13. Turret dialogue by Erbo · · Score: 4, Funny
    "Did you hear that?"

    "Someone is coming. Quick, tell them you see them."

    "Why? I don't see them."

    "It scares the hell out of people."

    "R-O-F-L! All right. I see you."

    "Ask her if the's still there."

    "Are you still there?...She didn't say anything."

    "No shit, Sherlock."

    "What should I say now?"

    "Come out, bitch!"

    http://www.youtube.com/watch?v=Uz5cl131KTk

    --
    Be who you are...and be it in style!
  14. Re:which goverment agency is going to grab this fi by Ocker3 · · Score: 2

    How much good-will does Valve generate from doing this (both from people who they buy them from, their friends, the people who see it in the lobby, their friends, and people who read about it, and Their friends), vs what piddling amounts of money they'd get for suing people, And the amount of good-will they'd Lose from doing that. Besides, who's making a living selling this kind of thing? It's not as if Valve is competing with these people and their products, it's an obvious synergy.