New Scientist recently reported on a snake-like robot that uses genetic algorithms to learn how to move. If a part of the robot is damaged, the GA process will find a new method to move about, if the damage is not too severe for it to be able to move at all.

This seems like just the task genetic algorithms will be good for solving.

Slashdot had an interesting discussion on the matter.

rf0 says: “I wonder if they could extend this to create a worm type robot that if it got cut in two it could still carry on. Say a brain in each segment”.

Very cool. If the whole mechanism was built in to every segment in the snake, parts as small as a couple of segments, could still go on and try to complete the tasks at hand. The snake might even use this deliberately, and split itself up from time to time, e.g. to cover more ground in a search operation or squeeze itself into tight corners.

thepacketmaster says: “While it is certainly natural for all living creatures to learn to cope with an injury, I think this is far more important for robots, since they don’t have a survival instinct. Humans will go to great lengths to avoid getting injured (well, most sane ones). However, robots will just do what they are told. While they may be told to avoid any hazards, they just don’t have that instinct that says I *really* don’t want to get my limb chopped off today.”

Good point and very philosophical. What exactly is a “survival instinct”, and how do you build one into a robot? Could a robot “think of” damaging itself in order to escape from a tight situation, like some animals do when in a trap?

Links:

• Another snake robot project at PARC: PolyBot
• Brilliant similar stuff by a guy named Mark Tilden.
• New York steam pipe fixing robot WISOR.