T Nation

Muscle Fiber Robots

One of my engineering classes led me to think about how mechanical limbs could be built using a ‘muscle fiber’ principle.

What if instead of motors (or other current means), we used bundles of many artificial fibers which:

(A) Contract, producing a force comparable to that in muscle fibers. (Piezo Electric materials contract and expand in the presence of electric fields, albeit very little. They’re used, for example, in jet-printers, where an electrical stimulus causes the piezoelectric material to contract, squeezing out the ink),


(B) consist of thin, resilient filaments which slide over each other (artificial Myosin-Actin filaments), each pair producing a small force. Bundle a myriad of such fibers into a bundle, and send a stimulus, causing the bundle to contract, producing a significant force. The type of stimulus (chemical, electrical) and materials for the fibers would have to be researched.

Or, yet still,

© Perhaps an easier idea to conceive: Produce fibers of an elastic material which elongates when squeeze, and return to their normal length when the ‘squeezing pressure’ is relieved. Encapsulate many such fibers such that, in order to produce force (from the stored elastic energy in the fibers), you need only relieve the pressure with which the capsule squeezes the elastic fibers.

In either case, each whole bundle of fibers would be stimulated - either electrically (piezoelectrics), chemically (the most likely scenario for fibers working on the Actin-Myosin principle), or physically (where you produce a mechanical force in the capsule, which squeezes the fibers allowing them to elongate. To produce force, you allow them to release their elastic energy by relieving pressure on the capsule, causing the fibers to shorten).

Think of mechanical counterparts for your upper arm and forearm. We would then attach each such bundle of fibers in such ways as to closely reproduce movements seen in our own limbs. By orienting each bundle very specifically you could now produce pronation/supination, flexion/extension. Perhaps 3 or 4 groups of such bundles (robotic ‘muscle groups’) woulc act on a specific limb and create any combination of the above types of movements.

The implications of reproducing good mechanical limb movement - comparable to that seen in humans or other animals - would be of tremendous use in robots whose purpose would be to enter hostile areas - such as those containing radioactivity for example.
Perhaps such mechanism would be more energy efficient.

I think this a good idea that might show some promise.

These are already being tested. I’ll try to search for the article.