Two Legged Robot Sets Speed Record

An anonymous reader writes "Researchers in Germany and Scotland have made the fastest two-legged robot yet (for its size) called RunBot. It is controlled by a simple program that mimics the way neurons control reflexes in humans and other animals. From the article: 'We wanted to show that a very simple system with a simple neuronal controller could walk in a natural manner - and fast,' says Florentin Wörgötter, from the University of Göttingen in Germany. The story also features couple of videos of RunBot in action."

Comments

videos

[mcguyver]

Two mpeg videos show RunBot (4.9M) walking at a steady speed and (15M) gradually learning to walk more rapidly.

Get'em while they're hot, ;)

Re:videos

[MustardMan]

If you're going to karma whore with video links, at least fucking coralize them.

Re:Um, does anyone else see the rod?

[rjstanford]

It can walk but can't maintain verticality? Is it there to stabilize it? That's pretty lame if they don't even have to worry about keeping its center of balance ... that's the hardest thing to figure out about fluid bipedal motion!

Probably just to stop the experiment needing a shitload of space without worrying about explicit turning. Of course, you could have just quoted the entire paragraph (two sentences) - bold face added:

RunBot currently walks around the edge of a circular room and is connected the centre of the room by a boom. But Wörgötter plans to develop a freestanding version next, and thinks it should be straightforward because the boom has only a small influence on its ability to walk.

Doesn't seem quite so problematic now, does it?

Re:How is it fast compared to a human?

[rrohbeck]

Taking my 34" inseam, 3.5 leg lengths/s would be 3.02 m/s or a 9-minute mile. Easy long run pace.

Gait generation, White paper

[m0nstr42]

This thing walks in a circle and is connected to a boom - it can't walk freely. All the legs have to worry about is front/back balance, and not side to side. Of course, making that obvious in the headline or summary would make the article seem much less interesting, and we couldn't have that, now could we?

Accepted. But that wasn't really the point of the research. If you RTFA and RTFWP (white paper), these guys are more interested in neuroscience. So what they did was design a simple mechanical system and a simple controller that both mimic the actual physical/physiological function of human legs. Balancing has alot to do with the structure of the foot and our ability to shift weight and is more of a dynamics issue. What these guys did was gait generation, which sounds simple (one foot after the other), but when you sit down and start trying to work out the details, its not so easy.

A similar (theoretical) study which actually addresses balancing (this time for insects with six legs):
J.E. Seipel, P. Holmes and R.J. Full (2004) Biological Cybernetics 91, 76-90. Dynamics and stability of insect locomotion: a hexapedal model for horizontal plane motions.