Skeleton Muscle Bot Brings I, Robot's Future One Step Closer

Eerily reminiscent of the design of Sonny and the other NS-5s in I, Robot, Kenshiro is the University of Tokyo's latest attempt to create a humanoid robot that accurately mimics human movement. And the researchers there believe the best way to build an artificial human is to simply copy our anatomy, particularly our muscular and skeletal systems.

Kenshiro uses some of the most advanced artificial muscles ever developed to move and walk — or shuffle, at least — in a manner that's surprisingly similar to how you and I get around. Standing at just over 150cm tall the robot closely resembles a 12-year-old boy, but weighs in at 50kg thanks to 160 different muscles bringing it to life.

And that's the biggest issue the researchers are currently dealing with: Kenshiro's weight. A full-sized human replicant would weigh in close to 100kg, which puts more strain on the muscle system, draws more power, and generally slows down its movements. So being chased by a horde of these robots is thankfully a nightmare that's not going to come true for at least a few more years. Phew! [YouTube via Automaton]


    just hurry up and make them, so they themselves can create a more perfect 2.0 model.

    I don't know. It seems to me that mimicking the human system is copying an arrangement that has plenty of flaws.
    Human physiology is far from perfection and a long way from being efficient.
    Don't get me wrong, for something that has evolved from a sea slug, human anatomy is astonishing to say the least, but there's plenty of room for improvement.
    Whatever these designers do, they really don't want to be copying the human spine.
    It's crap over the long run.
    Well, mine is at any rate.
    What wouldn't I give to be able to walk into a service center and have an updated, "new and improved" version fitted to replace the one I have now.

    This is pretty cool stuff just the same.

      The upright form also evolved that way due to our pre-human history of nomadic travel and hunting. Basing robot models on that paradigm is probably not the best, unless we want them to be efficient at those two tasks.

        The reason to build a robot that mimics human morphology is not just to make them efficient at nomadic travel or hunting, but to allow them to move and work in all the places that we humans live. The places we live are all designed with us in mind, from doorways, to chairs, paths stairs, buses cars etc (think Robocop vs Ed209)

        A humanoid robot could for instance, have gone into the fukushima reactor to work whereas a wheeled robot would have been stumped by the first set of stairs.

          Thats a good point, but the human form is not a perfect design in all cases. Perhaps a highly-articulated, multi-limbed centipede robot would be more effective in the fukushima reactor. Human-like robots are probably more important for activities related to interacting with real humans and realistically performing tasks real humans can/can't do.
          Working alongside human-like robots is a matter of time and there are many obstacles, including technological and philosophical, that humans must overcome to make the most of them.

      Really? can you show me anything that undergoes the stresses that things like joints and bones go through for 50, 60 or more years without degrading? How long will the bearings on a car last? How about the bearings used in excavators and the like? The human body is an amazing piece of macheinery and can teach us a lot.

    quick! patent the design of the human form, robot people!

Join the discussion!

Trending Stories Right Now