Three-legged dogs help robots learn to walkS

No matter how perfectly designed, robots still run the risk of suffering a serious accident with no chance of repair. To figure out how robots could adjust to losing a limb, scientists are examining the remarkable adaptations of three-legged dogs.

Dogs that lose a limb due to injury are forced to develop compensation strategies that allow them to walk with an odd number of legs. Since locomotion has developed in nature for animals with an even number of limbs, such an injury requires an innovative reorganization of the locomotive system, something roboticists hope to apply to robots that suffer their own limb-severing "injuries."

Martin Gross, a researcher at Germany's University of Jena, has had multiple three-legged dogs run on treadmills while infra-red cameras record their movements. What he and his team found is that losing a forelimb is a lot tougher for dogs to adjust to than losing a hind-limb. In the latter case, the two forelimbs are pretty much unaffected, acting just as they would if both hind-limbs were still in tact.

The loss of a forelimb, however, brings on a process known as gait compensation, where the remaining three limbs have to very carefully reorganize the way they work together to optimize the dog's movement. This is probably because of the way a dog's weight is distributed - after all, only the tail is behind the hind-limbs, but the forelimbs have to support not only the torso but the head as well.

[Society for Experimental Biology]