HRP-3 Promet II, Coming Soon to a Job Near You
The Mechatronic Systems division of Kawada Industries (what a name) has improved upon the amazing HRP-2 design by outfitting the robot worker of the future for harsh conditions, from the factory floor to the pouring rain.
Nicknamed Ma-Kun, the amazing new worker robot is dust-proof, water-proof, and has an improved balance system designed to navigate treacherous paths and slippery floors.
The HRP-3 Promet II, a product of Kawada, Kawasaki, and Japan’s AIST, was designed with the goal of developing a humanoid robot with sufficient skills to destroy all humans enter the workforce. Kawada hopes to court employers by offering the Promet II at $120,000 per robot by 2010.
After spending 5 years and $3.3 million to develop the robot, the group wisely hired anime-mecha designer Yutaka Izubuchi, known for his work on Gundam and Patlabor, to give Promet his “we got our development money’s worth” futuristic look.
Depending on its task, the robot can be configured to work autonomously or via wireless remote.
There’s a good write-up at Pink Tentacle, and a translated version of a very thorough Robot Watch article here.
UPDATE: Here’s a streaming version of the RobotWatch video files, posted by JapanProbe:
8 Comments so far
Leave a reply
please
send me, more information;picture-science technoloji data about hrp-3 please
turkey trabzon
23 .06.2007
Nice ! But, it was expected, indeed. Japan is apart of a history of humanity !
Quite impressive. It’s just a shame that they programmed him with the work ethic of a Spanish builder. He seems to think that if he puts it off long enough then those nuts will tighten themselves
Pretty amazing stuff. It’s cool to think about not even needing people to build a house (Not that this is even close to where we are technologically).
Also, Chris’s statement was pretty hilarious.
Given that AI isn’t that advanced at the moment, the robot’s usefulness on a factory floor is debatable. Why not use a wheeled robot or a conveyor belt or similar? And if operating via telemetry, there are stronger, more durable robots that could be used instead. But if it becomes more intelligent, then it would probably become genuinely useful… or evil
I agree that there is a little redundancy in creating a robot that mirrors human design when a wheel based design would be more appropriate but I think that is missing the point. The point of this (amazing) demonstration is that that the complex task of replicating the human form factor has been achieved and that building upon this much can be achieved. It is worth remembering that evolution designed humans successfully because of their adaptability and basing a robot design on this form gives a robot a base to perform many tasks rather than be specialised.
That’s excellent. Definitely thoughtful positions, both.
Wheeled or even installed robots are very efficient, and tend to be specialized in function. But they are dominant right now because the current state of AI lends itself well to specialization, and controlled environment.
And the human form, as evolved, is certainly very flexible and adaptable to unforeseen new tasks.
The thought I had on it this morning (while not particularly original) was that humanoid robot design was desirable not so much because the human form was the best of all possible designs, but because it’s completely ubiquitous at this point.
It’s desirable because humans have built most things to accommodate that form. From buildings to tools, vehicles and farm equipment, staircases and offices, we generally built the world to accommodate a roughly 5-ish foot tall bipedal walker with two hands.
If, as a robotics company, you can recreate this form reliably, it has the potential to become a drop-in replacement for all sorts of human tasks, without having to re-design the environment to accommodate the new worker.
What do you think?
Absolutely.
Humanoid form is inevitable, but these 1st-gen bots might not paint the whole picture in regards to the diversity(superiority) we might see in near future.
Specially with the whole Robotics vs. Bionics thing.
EAP’s looks interesting:
http://en.wikipedia.org/wiki/Electroactive_polymers
http://ndeaa.jpl.nasa.gov/nasa-nde/lommas/eap/EAP-armwrestling.htm
It would be interesting to see a more “soft body” approach that off course can be Humanoid, but at the same time have the diversity to take on other rigid or non-rigid forms as well, without the need of motors, joints etc all over.
http://engineering.tufts.edu/news/2007/MayBiomimetic.asp
The current state of AI might not be overwhelming, but the movement from narrow to a more broad,general framework in which it all can develop looks promising.
Perhaps in particular Goertzel’s Novamente approach.
Their take on developing it in regards to a simulated environment seems good and might help giving faster “proof of concept” and giving some meat on the thin bone of AGI.
http://www.8bithero.co.uk/2007/06/07/born-in-second-life/
In regards to the robots taking over peoples jobs:
http://marshallbrain.com/robotic-nation.htm is a nice read