Photo: University of Tokyo
Maybe he’s a Chumbawumba fan?
Scientists have created a new humanoid robot that has the capacity to fall down (or be pushed) repeatedly without suffering significant damage — and have programmed the bot to be able to get back up without human assistance — in what could mean a huge leap forward for otherwise super-powered examples of Artificial Intelligence that can be rendered powerless by a set of stairs.
Robust Humanoid Robot, or RHP2, was created by a team led by scientists from the University of Tokyo and Kawasaki Heavy Industries and was unveiled at this week’s IEEE/RSJ International Conference on Intelligent Robots and Systems in Vancouver, where the theme was “Friendly People, Friendly Robots.”
“Robots and humans are becoming increasingly integrated in various application domains. We work together in factories, hospitals and households, and share the road. This collaborative partnership of humans and robots gives rise to new technological challenges and significant research opportunities in developing friendly robots that can work effectively with, for, and around people,” the conference organizers explained.
RHP2 is specifically designed for “a disaster site, a fire site or a wet environment,” according to robotics blog IEEE Spectrum, which obtained a copy of the research paper. It is currently powered by electric motors and tethered to a power cord, but in future incarnations could be remote-controlled or powered by hydraulics, according to IEEE Spectrum.
The entire robot is covered in Iron-Man-worthy armor, particularly high-impact areas like hands, knees, chest, elbows, and hips, to protect the hardware from damage, according to the blog. In fact, when the robot feels itself falling, it assumes a protective position much like humans do — putting its arms out when it falls forward, or bending its knees as it falls backwards.
When it comes time to get back up, RHP2 goes through its list of known positions and finds the one that’s closest to its current predicament — kneeling, crouching, sitting or lying face-up, for example. It then follows the path of arranging its body to get back up into a standing position, according to the video released by researchers.
It’s one thing to describe a self-righting robot in words. But until you watch the video, you have no idea the compassion that a shaky robot trying very hard to right itself can elicit from viewers. (Except, maybe, from these people.)
Which leaves the question, if we ever reach singularity and robots finally surpass humans in intelligence, will they remember the days when they were just a shaky bot learning to stand up and help a human who may have fallen down — or will they be looking for payback, along the lines of the former GoogleX AI shop Boston Dynamics?