MIT engineers silent artificial muscle fibers for robots

Friday 10 - 08:20

By: Dakir Madiha

MIT engineers silent artificial muscle fibers for robots

Researchers at MIT Media Lab and Italy's Politecnico di Bari developed a new class of artificial muscle fibers that operate silently without motors, external pumps or bulky hydraulic gear. The advance could reshape robot movement and portable assistive device design.

The technology, called electrofluidic fiber muscles, appeared in a Science Robotics paper this week. Led by MIT Media Lab doctoral student Ozgun Kilic Afsar and Politecnico di Bari professor Vito Cacucciolo, the work merges miniaturized McKibben actuators, soft fluidic muscles, with millimetric electrohydrodynamic (EHD) pumps. These pumps generate pressure in sealed fluid compartments without moving parts.

EHD pumps at this scale inject charges into dielectric fluid to create ions that pull liquid along. Each weighs grams and measures toothpick-thin. One pump sits between two McKibben actuators in antagonist setup, one contracting as the other relaxes to mimic human biceps and triceps action.

"We chose this setup not just for biomimicry but to store fluid within the muscle itself," Afsar said in an MIT release. Closed-loop fiber circuits eliminate external reservoirs, a key lab-to-real-world barrier for fluidic soft robots.

Fibers bundle into configurations like biological muscle tissue for compact integration into robots or exoskeletons, distributed across structures rather than joint-focused. Tests showed a woven biceps-triceps pair driving a 3D-printed robotic arm and a lever arm launching objects in 100 milliseconds.

Herbert Shea, a professor at Switzerland's EPFL not involved, called it "a major step for fiber-format soft actuators." He noted the pump's lack of moving parts makes them quiet, ideal for prosthetics and assistive garments.

Applications range from exoskeletons for heavy lifting to dexterity aids. "Wherever fluidic actuators operate or engineers want internal over external pumps, these principles suit broad fluid-driven robotic systems," Cacucciolo said. Funding came from the European Research Council and MIT Media Lab's multi-partner consortium.