Energy-Efficient Robots

By | February 18, 2005
4 Comments

An interesting breakthrough, robots that expend about the sae amount of energy
getting themselves around as human beings do:

A
trio of androids that amble along with exceptional power efficiency and "instinctive"
co-ordination were unveiled for the first time on Thursday.

The three mechanical bipeds, built by researchers from Cornell University,
the University of Michigan and Massachusetts Institute of Technology in the
US and Delft University in the Netherlands, respectively, walked along in
an amiable, if somewhat awkward fashion.

"Our robots demonstrate that utilising the natural dynamics of the body
can make robots much more efficient," says Steve Collins, a member of
the team from the University of Michigan. "For any autonomous robots
to be practical, they must be energetically efficient."

Contrast these energy-efficient droids with Honda’s Asimo they’re about 10
times more efficient. That’s huge. But I imgaine they’re still quite expensive
to build. Still, it won’t be long before somebody finds something useful to
do with these kinds of robots. Then two things will happen:

1. The cost will go way down.

2. They will become much more energy efficient than human beings.

For better or for worse, those developments will mark the beginning of the
end of (what’s left) of the manual labor market. After that, we’re only a few
steps away from a world that Dr.
A. himself would have found strangely familiar.

via Kurzweil AI

  • https://www.blog.speculist.com Stephen Gordon

    Something I’ve often wondered – maybe some software guru can help with this – how are the three laws going to be written into robots?

    In these early days of robotics it will have to be simpler than some overiding positronic routine or whatever. These early robots aren’t the sentient machines that Asimov envisoned. It might just be simple common sense safety precautions.

    I saw a program the other day about industrial robots. One large car-building robot was working fast and furiously within a chain-link cage. The moment the technition unlatched the gate to the cage, all work ceased. That big machine could hurt somebody had it continued working with someone inside.

    That industrial safeguard was “Law One” in action.

    Here’s the laws:

    http://www.auburn.edu/~vestmon/robotics.html

    1. A robot may not injure a human being or, through inaction, allow a human being to come to harm.

    2. A robot must obey orders given it by human beings except where such orders would conflict with the First Law.

    3. A robot must protect its own existence as long as such protection does not conflict with the First or Second Law.

  • https://www.blog.speculist.com Phil Bowermaster

    Michael Anissimov says that the three laws are too simplisitc ever to be coded into robots.

  • Engineer-Poet

    It may be possible to become more energy-efficient than humans using hopping motions, but it’s hard to see how order-of-magnitude improvements could be made without going to wheels.

    I don’t see why energy efficiency is all that important, save for the military; most robots in industrial or domestic settings will have power readily available, and their engineering can concentrate on other useful attributes.

    (Incidentally, the preview window claims that I am no longer logged in, and won’t let me post without going back. Strange.)

  • Karl Hallowell

    Energy is cheap. That’s not the factor that’s going to cause humans to be obselete. When it becomes cheaper to use a robot than to hire sufficient humans, then that’ll be the time to switch. Needless to say, that point arrived already for a number of applications. General manual labor is still some distance away.

    Steve, much of the analysis on “3 Laws Unsafe” isn’t that good. I have yet to see anything there that tops Asimov’s idea. Simple even simplistic beats complex systems that you can’t guarantee. And frankly a lot of the arguments there suffer from simplistic reasoning just as well. For example, Michael Anissimov advocates allowing robots to evolve uncontrolled except for initial instructions. Even assuming that the robots couldn’t evolve away the initial instructions, it’s still a tremendous gamble that the unintended consequences won’t be too dire.