A printed hinge turns a pinch, bend or squeeze into a control signal

A lamp changes its brightness when someone pinches its shade. A glove records a hand gripping blocks and passes the movements to a robot hand. A shark-shaped game controller bends its tail to register a player's gesture. Researchers at MIT and Tianjin University built these devices around a flexible printed hinge that measures how it bends, integrating the sensor into the joint itself.[1]

The system, called X-Hinges, was described by MIT on 22 October 2026 ahead of an early-November human-computer interaction conference presentation. It offers design software that embeds sensing elements while fabricating an object by multi-material 3D printing. This matters to AI and robotics for a physical reason. An intelligent system can act only on signals that describe what is happening in the world, and many everyday objects have no useful signal at all. X-Hinges explores what changes when sensing becomes part of the object's structure, not an accessory bolted on afterwards.[1]

The trouble with printing a sensor and its wire together

The idea sounds straightforward: print flexible plastic, put an electrically conductive path across the bend and detect the change in resistance as the joint moves. But a sensor must be connected to a circuit. If its sensing element and the conductive traces use the same electrical material, the traces themselves can respond to strain. Motion in an unintended place becomes mixed with the signal from the part scientists actually wanted to measure. The problem grows when a hinge can bend in several directions at once. Separating one direction from the others then becomes a practical signal-processing problem, not just a design challenge.[1]

The researchers' printing approach combines a nonconductive structural body with conductive filaments selected for different electrical roles. One conductor supplies the sensitive resistive element; another forms connecting traces designed to interfere less. Their software helps position those features automatically inside a compliant hinge. Users choose the geometry and intended motion, including axial compression and bending across two other directions. In the associated research paper, the authors describe three degrees of freedom with dedicated sensing configurations and data-driven regression models to estimate movement from the changing electrical signals.[1]

Four prototypes, four different design problems

In the glove prototype, sensors measure finger bending and use those readings to guide corresponding movements of a robotic hand. The interface provides physical guidance for grasping, making movement capture one potential use of the printed sensing method.[1]

The origami-like lamp is a gentler example. Its flexible structure senses a user's pinch and adjusts light intensity. Instead of relying on a conventional switch, a designer can turn deformation of the object itself into an input. A sensing platform showed that touch signals could distinguish an object placed on it, illustrated by a banana. A playful controller used a shark tail as a joystick. The variations are important: they show that the core achievement is neither lighting nor fruit identification. It is a reusable fabrication method for different geometries and interactions.[1]

Every print is a little different

Small differences in conductive material placement can change the readings from separately printed devices. The researchers used computer-vision-based recalibration to align them. Calibration and durability will therefore help determine which applications can use the method beyond a carefully prepared prototype.[1]

The work covers prototype fabrication and measurement. Integrating actuation—the ability to move—alongside sensing is a further development the researchers propose.[1]

What happens when objects become interfaces?

The deeper possibility is a change to the cost of physical interaction. Many objects are designed as passive things: a glove fits a hand; a lamp holds a bulb; a toy bends. Turning part of a physical structure into a continuous input channel could let each object describe its own deformation to an external system. Such signals might help robots learn hand movements, give designers new ways to create interfaces, or help researchers collect motion data without a bulky collection of separate sensors.[1]

A lamp becomes a more expressive control when the shade itself can measure a pinch. A glove becomes a motion interface when its joints can report their bending. X-Hinges makes those interactions part of the object’s construction.[1]

Report an error or suggest a correction