Slashdot Mirror


Robots Do The Darndest Things

alito writes "15 years ago they couldn't get them to walk, now they are rollerskating (video). Read more about the 2004 Intelligent Robotics and Systems conference in this New Scientist article, and at the conference's site. Also shown at IROS, a childbirth simulator for obstetricians, a capsule that crawls through your intestines, and a 3-mm long swimming robot. (No, I don't get paid by New Scientist.)"

12 of 145 comments (clear)

  1. Re:what's the point? by horrens · · Score: 5, Insightful

    actually it's a pracitce in balance
    if you ever have tried rollerskating then you should know that it requires quite good balance and body control

  2. Rollerskating robots.... by Anonymous Coward · · Score: 5, Insightful

    I think you may be missing the point. Rollerskating, running, walking, what have you, are important fundamental tasks that must be worked out before you will *have* a robot that can do something advanced like wash your dishes or walk you dog. If it can't rollerskate with perfection how do you expect a robot to perfrom a mundane task competently. I for one welcome rollerskating robots.

  3. We're headed for trouble by beware+of+the+robot · · Score: 4, Insightful

    In my opinion, it is not a question of *when* we'll get trouble from the robots we're going to create, but what we're going to do about it.

    We might as well start planning right now. The article in the posting, as well as numerous other reports, show that the robot mechanics is getting better and better. What is lacking, is some real AI. I think that within 30-50 years, if not before, this "problem" will also be solved.

    That's when trouble starts. As Isaac Asimov shows in his literary "experiments" with the three laws of robotics, even *with* benign top priority imperatives NOT TO HARM humans, we may not be safe.

    Given the mechanics and the AI, how will robots become a threat to us? Imagine a scientist in a robot laboratory sometime in the future. He/she has all the parts needed to produce the ultimate robot: agile and completely autonomous. I think the temptation will be too much for *any* person. It will be impossible to refrain from releasing a completely free and autonomous robot into society to see what happens. Imagine the excitement: This is comparable to a second creation; it is almost like being God.

    Naturally, the robot will have a strong need for self preservation. So it will start to secure land, natural resources, labour, spare parts, factories, and so on, and build other robots and societies to fulfill these and other purposes.

    This is when conflicts will start. Wars often start as a result of a disagreement over natural resources or land.

    Ideas, anyone?

    1. Re:We're headed for trouble by groomed · · Score: 4, Insightful

      This is wrong on so many levels.

      First, the practical problems involved in creating a robot that can autonomously participate in human society is far from being solved. People have been saying for over 40 years that the solution was 30 to 50 years in the future. I wouldn't expect it to arrive in a 1000 years, if ever. And yes, I study AI.

      Second, as Sartre observed, "Hell is other people". A single super robot on the loose is no match for our puny weapons. To be effective, he'd have to enlist a following. But since he's alone, those followers would have to be recruited amongst humans. But how on Earth is that ever going to happen?

      Third, you make it all sound like a video game. "Secure land, natural resources, labour...". You took a page straight out of Civilization, there.

      Fourth, there are a lot more pressing issues to worry about than what happens when a breed of superintelligent robots wants to dominate the planet.

    2. Re:We're headed for trouble by Anonymous Coward · · Score: 1, Insightful

      If your robot has a strong need for self preservation, why would this automatically result in robot societies and inevitable conflict?

      Surely there's another way any robot could preserve itself: get a job. I'm serious! Assuming artificially intelligent robots aren't all-powerful laser-shooting giant killing machines, there's no reason they couldn't be as punishable by the legal system as us humans.

      I mean, last time I looked, us humans had a pretty strong need for self preservation, and we don't generally go round trying to kill all the other humans. Well, most of us don't. And those that do generally get locked up. Or go and hide in a cave in Afghanistan.

  4. When we think "robot", we think "android". by reporter · · Score: 3, Insightful
    When we think "robot", we invariably think "android": a mechanical device that appears like a human being and mimics some of its actions. We are fascinated by androids for the very same reason that we are fascinated by apes. They look like us.

    The attraction for androids is only skin deep. Today's androids are just a mass of wires. Getting a robot to walk, shake hands, play chess, etc. is substantially different from a sentient machine.

    Sentience impresses me, but a mechanical shaking hand does not.

  5. That robot is NOT rollerskating by Scrameustache · · Score: 5, Insightful

    Its walking akwardly while wearing wheels on its feet.

    Which, to anybody who's learned to skate, is of course remembered fondly as those first steps before you learned why they made you wear those uncomfortable wrist guards.

    Its a nice little robotic achievement, but its not skating.
    Actual skating would involve a phase of sliding along between "steps".

    --

    You can't take the sky from me...

    1. Re:That robot is NOT rollerskating by dhudson0001 · · Score: 2, Insightful

      I'm not responding directly to anyone in particular, but to the whole coalition of /. folks running around this topic who imho have failed to recognise this moment for what it is..I really just couldn't read any further through the thread w/out responding... Just everyone slow down for a moment....We have a Robot...and it IS skating around on a table like a very agile small human might, and It doesn't take a extreme leap of vision to see what is up next...I don't know about the rest of you..but watching that video all I could think about is how smooth that little guy looked...perhaps some of us should consider watching it again (post-s/d effect)and notice there IS a marked step_slide movement going on..also check out the nice sway of the "hip" joint-whatever... Really people....this whole thread makes me wonder when did we become so unappreciative/spoiled that we are unable to enjoy this very cool moment without the high-faulting critical attitude..Look..I for one have no doubt in 15 years I will look at the bot who cleans my house,washes my clothes..etc..and recall fondly the first time that I saw ASIMOV WALK...then I will prob remember that darned rollerskating bot next.... and I will enjoy the satisfaction of knowing that I didn't blindly pass up either page in robotic-ancestery/history with nothing more than a sideways glance and a off-colored comment about how silly the whole thing was. Some seem to want to forget how truly extraordinary these times we live in are...

  6. Personally, I'm not that impressed by kai.chan · · Score: 3, Insightful

    Making the QRIO skate is not really a big technological achievement considering that the leg motions have already been pre-programmed to get the robot to walk. It is really a matter of calculating the weights and trying to balance the robot while moving the legs to propel the robot.

    Now, if they can program a neural network that changes the leg and arm movements in relation to a physics model, and have the robot learn how to walk and skate by trail-and-error, then I'd be more impressed. THAT is what we should aim for nowadays.

  7. Re:what's the point? by Anonymous Coward · · Score: 1, Insightful

    but what about making ones that can help ALL people?

    I disagree. We don't want robots to go the way of cellphones, where they can make calls, take photos, play mp3s, etc, all in a mediocre way. I'd prefer to have an army of small robot each specializing in a handful of tasks. This way, each robot can be built well by a specialist in the area, and when the shopping robots breaks I can still use the cleaning robot, or upgrade the sex robot without upgrading my singing robot...

  8. Re:what's the point? by rebelcool · · Score: 4, Insightful

    The point is to demonstrate advancements in balanced mobility. Its very difficult to get a robot to walk well with anything resembling 'legs'. Roller skates introduce lots of unpredictable uncertainty into walking, and to have a control system fast enough to detect and adapt to the changes is very impressive.

    As most things in our world are built around our type of mobility (legs), an autonomous real world robot will interact with us and our world far better if it emulates our system of mobility.

    Thats the point.

    --

    -

  9. Killer autonomous robots will be intentional by Anonymous Coward · · Score: 1, Insightful

    Look at all the computer virii, worms, etc. released over the past 20 years. None of these were an "accident". Every single mallicous program ever created has been an intentional act.

    So considering robots and AI, if intelligent robots decide to start killing and take over the world it will be by human design before it's ever a result of robotic nature and/or independant decision. Some mallicious person will attempt to build a robot that has the skills necessary to reproduce and cause as much damange as possible.

    Sounds very similar to the function of a computer virus today right?