Insect-scale robots need actuators that are small, fast, strong, and low-power. Existing approaches trade between these requirements. Piezoelectric actuators are fast and precise but produce tiny displacements and require high voltages. Dielectric elastomers stretch far but need kilovolt fields. Shape-memory alloys are strong but slow. Each technology optimizes one axis at the cost of another.
Xu, Cao, and Jiang at Westlake University built an actuator that borrows from two biological systems at once. The primary inspiration is muscle contraction: an elastomeric polymer structured to contract when an embedded coil generates a magnetic field, producing about 210 newtons per kilogram of actuator mass — comparable to biological muscle — with up to 60% contraction ratio and response at 60 hertz, all at less than 4 volts.
The second inspiration is subtler. Mollusk catch muscles can lock into a contracted position without continuous energy expenditure, holding shells closed for hours while the animal rests. The actuator mimics this with a bistable mechanism: it can snap into a contracted state and hold it without power, then snap back when the field reverses. This means the robot holds its leg position at zero energy cost during stance phases, spending power only on transitions.
The combination produces insect-scale robots that walk, climb, and carry loads while running on small onboard batteries — genuine autonomy rather than tethered operation. The low voltage means the power electronics shrink to match the robot's scale, eliminating one of the persistent bottlenecks in miniature robotics.
The through-claim is about the distance between the inspiration and the mechanism. Muscle contraction in animals uses actin-myosin molecular motors, ATP hydrolysis, calcium signaling — none of which the actuator replicates. Mollusk catch muscles use paramyosin filaments that lock mechanically — the actuator uses bistable elastomers instead. The biology is not copied. What is copied is the functional architecture: a high-force, high-strain primary actuator paired with a zero-power holding state. The robot does not have muscles. It has something that does what muscles do, using completely different physics. The blueprint is not the mechanism but the specification — what the mechanism must achieve, not how the original achieves it. The robot borrows the latch, not the lock.