Remote Control Robotic Snakes

0xdeaddeaf writes "Check out this site on remote controlled robtoic snakes from Dr. Gavin Miller. He's been working on a realistic moving robotic snake for a number of years and has posted several videos of his prototypes that span from S1 in 1987 to S5 in 1999. The snakes are self contained with onboard computer, battries, receiver, and locomotion system. The video of S3 shows they can move like sidewinders and S5 does indeed look extremely realistic. Put a skin on these things and there is no doubt you fool a lot of people if the motors are quiet. The New Scientist has an additional article that explains how their movement is performed. "

Comments

Giving a man a fish vs. teaching him how to fish

[Kaufmann]

Although this is a pretty cool engineering feat and will probably eventually become useful, I don't consider it to be very meaningful in the Big Scheme of things. IMNSHO, biomimetics is the result of people getting what is essentially the wrong message from nature.

What I mean is: people look at, e.g., a snake moving around gracefully, and they say "Let's replicate that kind of movement! It'll make for very useful robots!" So they go and study snake physiology for years on end, eventually mapping down the entire motion mechanism. Then comes the hard part: they try to build an artificial motion mechanism that works on the same principles. This proves incredibly frustrating, and it takes a Long Long Time and plenty of irritating compromises before someone comes out with a workable design - which will likely resemble the original in very few aspects. In effect, they're just trying to reverse-engineer a finished product [1], knowing only assembler code and nothing of high-level programming.

Now, I think a much better way to do things is to notice, instead, how incredible it is that such a powerful and versatile, while buggy, mechanism has arisen from just a few organic molecules without external intervention [2], and to concentrate on learning how the evolutionary process works - and to try and mimick /that/. This may not (in fact, it probably will not) produce motion mechanisms that are identical to snakes', but it will at the very least produce systems that are equivalently adapt at moving in the same kind of environment. Not only that, it'll also provide us with an endless source of new ideas for systems and of insight as to how to design better ones. In effect, instead of copying nature's mechanisms, we are learning to program for ourselves.

Of course, I'm aware that many steps have already been taken in this direction (the GA-designed Lego Bridge comes to mind), but GA-based tech is nowhere near as popular a research field as traditional biomimetics.


[1] Yes, I know, nothing in nature is a "finished product". I just used the term for the purpose of analogy.

[2] ... which is what is generally thought to be most likely to have happened. Before you try to assert that there was any kind of external (external to the environment - i.e., supernatural) intervention, let me tell you that extraordinary claims require extraordinary evidence, and until I see some of it, I'm sticking with the strictly-natural theory.

Cool!

[Anonymous Coward]

I hope they have made a Python interface for it...

duh

[mwalker]

whatever. everyone knows dr. doom invented the robotic remote-controlled snake. it sucks though because spiderman figured out how to modify his webslinger to stick to the veno-bot's poisonous scales well before veno-bot could send the city into chaos.

please do your research people.

Copper snake?

[pnevares]

I have my robot snake running on a Coppermine and it won't wake up. Any ideas? =)


Pablo Nevares, "the freshmaker".

Re:What happen to S4?

[Black Parrot]

They had an infinite loop in the control code, so it swallowed its tail and disappeared.

--
It's October 6th. Where's W2K? Over the horizon again, eh?

Robotics

[Ex Machina]

Supposing I wanted to make some computer controlled robotics, what would be the best way to start? Has anyone found a Mindstorms (tm) like system but for controlling stuff hooked up to a computer through a C program. Like uhh
#include

int main() {

robot arm(arm, 12, 120, 3, 2);
/* 12 volts 120 hertz AC,
3 axis of movement, 2 other motors
*/

while (!arm.is_touching())
arm.move(1,1);
// (move the arm 1 degree
// along the x axis until it touches)
arm.motor(2, +32767);
// crush it
return 0;
}

Do you think one of these could eat an AIBO?

[1010011010]

"When robot snakes attack," tonight on FOX!

A family of four lost their Aibo last night when, tragically, a loose robot snake ate it. Robot snakes typically need to eat only one Aibo a month to survive, and, once satiated, return to their nests.