Engineers at the University of Pennsylvania have built an artificial muscle that can make fine, steady movements, the kind needed to move a finger or hold an eye in place. Popular Science reported that the team tested it on a robotic hand that struck piano keys with different levels of force and on a pair of robotic eyes that shifted their gaze on demand.
The work was published this week in the journal Advanced Functional Materials. It is a laboratory demonstration, not a product, and the researchers describe their next steps as combining more of these muscles into larger systems.
According to Popular Science, most large robots, humanoid ones included, depend on bulky, rigid motors and components. That works for walking and lifting. It works far less well for the delicate muscles that move human eyes and fingers, which call for flexibility, fine control and enough strength to survive repeated use.
The article notes that tasks that look easy to people are often the hardest for machines. Robots have performed dance routines and acrobatics for years, it said, yet they still struggle with folding laundry or holding a gaze.
Shu Yang, a University of Pennsylvania engineer and study co-author, said in a statement that soft robotic muscles must be strong and flexible at once, and that those properties are rarely found in a single material.
The team built each fiber from three parts. The main material is a liquid crystal elastomer, a rubber-like substance that acts as the muscle by contracting and stretching. Graphene is added to make the material tougher and a little softer, so it holds up when opposing muscles pull against it. An ultrathin copper coil runs through the middle.
The coil solves a specific problem. When the elastomer is heated, it stretches, but it then keeps contracting and slowly creeps back toward its original length as it cools. That drift would ruin a light, steady key press or a fixed gaze.
The researchers used the coil as both a heater and a sensor. Study co-author and Ph.D. student Xuelin Wu said the team tracked the coil's electrical resistance as a feedback signal and adjusted the heating power in real time to keep the resistance near a target value. Because one part does both jobs, no separate sensor is needed.
In the first test, the muscles were fitted into a 3D-printed robot hand built to mimic finger movement. The fingers pressed piano keys at different levels of force and speed, and sensors under the keys recorded how hard each note was played. Popular Science described the result as no Mozart, but a working demonstration.
In the second test, the team mounted a pair of robotic eyeballs on freely rotating ball joints inside 3D-printed shells. Muscle fibers were attached around the edges of opposing pairs, imitating the way eye muscles pull against each other. The eyes shifted their gaze on demand with controlled, repeatable movements, according to the report.
For most people, the near-term effect is limited. The study shows a design for small, precise artificial muscles that could one day replace some rigid motors in robots that need to move with subtlety, such as in faces and hands. Popular Science says the long-term goal is robots that look and feel more like the real thing, and the outlet also remarked that the idea is unsettling to some.
The research is part of a growing field of labs and companies working on artificial muscle systems. Popular Science pointed to Polish startup Clone Robotics, which in 2025 released a video of a faceless, human-sized robot hanging from a ceiling and writhing. The company says the movements were driven by synthetic muscles and a hydraulic system, and calls the machine the world's first musculoskeletal android. Popular Science said the video was likely more a proof of concept than a working prototype.
The Pennsylvania team says its next step is joining more of these muscles into larger, integrated systems, where dozens of small muscles must cooperate to produce more natural expressions. Later, the researchers are interested in using AI computer vision to coordinate such systems.
Ziyun Zhang, a University of Pennsylvania engineer and co-author, said he is excited about using the approach to build robots that move with the subtlety and expressiveness of human faces.
Around Prague, based on reporting by Popular Science