An elephant’s trunk is one of the marvels of the animal kingdom. With up to 40,000 muscles, the trunk is powerful enough to lift a 600-pound log and delicate enough to crack a peanut shell without damaging the nut.
That kind of versatility has inspired German engineering firm Festo to design a bionic elephant trunk made from 3D-printed segments and controlled by a system of pneumatic muscles. Led by Jochen Steil, an intelligent systems engineer at Bielefeld University, the firm the wants to mimic the real trunk’s dexterity to push, pull, reach and grab, and apply it to industrial robots that can perform jobs like changing light bulbs or picking apples.
However, the the robot trunk was not designed with precision control software. Instead, engineers used a method called “goal babbling,” a process of trial and error learning, similar to how infants discover their muscles. As the robot goes about tasks, it records the tiny pressure adjustments made in the pneumatic tubes operating the artificial muscles. This “learning” creates a map that relays the trunk’s exact position to calibrate the pressure in each tube. As engineers manually manipulate the trunk into position, such techniques allow the bionic trunk to be trained on command.
While the method may feel awkward at first, New Scientist’s Paul Marks reports that, just like humans learn, the bionic elephant trunk eventually gets the hang of it. Marks even got to take the trunk for a test drive.
“As I move the bionic trunk in Steil’s lab into different positions it initially resists, but then yields and follows my movement,” he writes. “The next time I try to push it to the same spot, it moves easily, because the behavior has been learned. The robot now has muscle memory — which makes it seem even more alive.”