Shape-memory wire simulates muscle in high-precision artificial hand

Team member Filomena Simone with the prototype hand (Photo: Oliver Dietze)

Whether they're on robots or amputees, artificial hands tend to be rather complex mechanisms, incorporating numerous motor-driven cables. Engineers from Germany's Saarland University, however, have taken a different approach with their hand. It moves its fingers via shape-memory nickel-titanium alloy wires, bundled together to perform intricate tasks by working like natural muscle fibers.

The individual wires are so thin that even when a number of them are wound together, the resulting bundle is still only about as thick as a cotton thread – but it has the tensile strength of a much thicker metal wire.

When the nickel-titanium alloy wire bundles are heated by conducting a supplied electrical current, the lattice-like internal structure of the alloy responds by contracting, causing the wires themselves to get shorter. Once they cool down, the wires return to their original relaxed state. That's why bundles of tiny wires were used, instead of single larger wires – the increased surface area of the bundles allows the wires to cool rapidly, once the electrical current is shut off.

On the prototype hand, each finger has one wire bundle running up the length of its top surface, and another going along its underside. These play the parts of the extensor and flexor muscles, respectively, allowing each finger to both curl up and open with high tensile force.

Additionally, by measuring the electrical resistance of each wire bundle, the semiconductor chip that controls the hand is able to ascertain the position of each finger in real time. This allows for very precise movements, and the ability to quickly respond to commands.

Although the hand is still in the developmental stage, the researchers have stated that it should result in a finished product that's lighter, quieter, more flexible and less expensive than current artificial hands.

Led by Prof. Stefan Seelecke, the Saarland team is now working on modeling hand movement patterns, and looking for an industrial partner. The prototype can be seen next month at the Hannover Messe industrial fair.

Source: Saarland University (German)

Top stories

Recommended for you

Latest in Materials

Editors Choice