What engineers are trying to copy
Natural muscle is a remarkable actuator. It is soft, quiet, self-repairing and able to produce force over a wide range of speeds, and it contracts when it receives a signal. Most machines use something different: an electric motor spinning a shaft, then gears or linkages to turn that rotation into useful movement. That works well in many places, but it is rigid and heavy for some jobs, especially wearable devices, small robots and prosthetics.
The goal behind “artificial muscle” is therefore a material or device that shortens, bends or twists on command, ideally with a muscle’s mix of force, stroke, speed and efficiency. No single technology matches all of these at once. The history is a series of partial answers, each strong in some respects and weak in others.
Pneumatic tubes in the 1950s
The best-known early design is the pneumatic artificial muscle often named after the physicist Joseph L. McKibben, who is generally said to have developed it in the 1950s for powered braces for polio patients. The design is simple: a rubber tube is wrapped in a braided sleeve. When air pressure inflates the tube, the braid bulges outward and shortens, so the whole device pulls like a muscle.
The McKibben muscle can produce considerable force for its weight, and the design has been revived and refined many times, including in some robot arms. Its weaknesses are practical. It needs a compressor or air supply, it responds somewhat slowly, and its force is not easy to control precisely. Those limits keep it from being a universal replacement for the motor.
Shape memory alloys
Another thread began with metals that remember a shape. The nickel-titanium alloy known as nitinol was reported around 1960 by researchers at a U.S. Navy laboratory, and it became the best-known shape memory alloy. When a deformed wire is heated, whether by electric current or otherwise, it returns to its original shape and shortens, producing a strong pull.
The effect is compact and forceful, and thin wires have been used as small actuators. The trouble is speed and efficiency. Heating is quick but cooling is slow, so repeated cycles are slow, and a lot of energy turns into heat. Lifetime under repeated cycling is also a concern in demanding uses.
Polymers and gels
A third line of work looked at materials that change shape when a chemical or electrical signal is applied. Collagen fibers were shown in the late 1940s to contract and relax in response to changing chemistry, an early demonstration that a non-living material could do something muscle-like. Later researchers developed polymer gels, ionic polymer-metal composites, conductive polymers and dielectric elastomers, which squeeze and stretch when a voltage is applied across them. These electroactive polymers are a large subject, and some of the ideas in them show up alongside shape-changing materials studied in 4D printing.
Interest in these soft materials peaked around the late 1990s and 2000s. Some can produce large changes in shape, and many are lightweight and quiet. But they commonly need high voltages or are slow, weak or short-lived in comparison with natural muscle, and few have moved beyond specialized uses.
Fibers and yarns
Around 2009 and afterward, groups including one led by Ray Baughman at the University of Texas at Dallas described yarns made from carbon nanotubes that twisted or contracted when stimulated. In 2014 the same group reported that ordinary nylon sewing thread or fishing line, twisted and coiled into a spring-like shape, shortened noticeably when heated. Because such fibers are cheap and are essentially derived from everyday synthetic fibers, the result attracted a lot of attention. Like shape memory alloys, these fibers depend on heating and cooling, so speed and efficiency remain challenges.
Soft robots and the present
Many present-day systems combine ideas. Soft robots often use pneumatic or fluid-driven chambers, which are close relatives of the old McKibben tube. Wearable assistance devices use cables, pumps or fabric actuators. Research continues on hydraulically amplified actuators, electrically driven polymer designs and materials made partly from living muscle cells, although the last is still at an experimental stage.
Why the motor still dominates
Considering the whole history, artificial muscle is better understood as a family of tools than as one invention. Electric motors improved so steadily, and lithium-ion batteries and control electronics improved with them, that motors were a hard target to beat. Artificial muscles therefore tend to win where softness, light weight or silent operation matter more than raw efficiency, such as in some grippers, exoskeleton add-ons and experimental prosthetic limbs.
Claims in news coverage that a new material will “outperform human muscle” usually refer to one measure, such as strain or force per weight, rather than to every property at once. Reading such claims with that in mind is a good guide to how much real progress has been made.