If developing next-gen robotics is your thing, Hiro's your man.
A couple of years ago, I introduced you to Hiro, a QNX-based robot designed for research and teaching programs in university labs. Even if you didn't read about Hiro here, he may still seem familiar, what with his appearances on Engadget, übergizmo, and other über-popular sites.
Kawada Industries, the company that created Hiro, describes him as a starter set for research into humanoid robots. To that end, Hiro comes equipped with a stereo vision camera, speech recognition, hands with 15 degrees of freedom, hand-mounted cameras, and a repeat positioning accuracy of less than 20 micrometers — that's 20 one-thousandths of a millimeter.
Since my last post, Kawada has uploaded some videos to demonstrate Hiro's chops. For instance, here's a clip showing how he has all the right moves:
And here's a clip showing how he can listen to voice commands:
If Hiro's role is to serve as a platform for next-gen robotics, he is succeeding. Recently, Osamu Hasegawa, a professor at the Tokyo Insitute of Technology, used Hiro as the basis for a new "thinking" robot. The robot, also dubbed Hiro (confusing, I know), employs a Self-Organizing Incremental Neural Network algorithm to adapt to its environment and learn new tasks.
For instance, in this video, a researcher asks Hiro to pour a cup of water. Hiro has never done this before, but figures out how to do it. That's some algorithm!
For more information on Hiro and his manufacturer, Kawada industries, click here.