MIT's X-Hinges prints motion sensing into flexible joints for gloves, lamps and controllers

A flexible joint usually supplies movement; a separate sensor measures it. Researchers at MIT and Tianjin University have combined the two in X-Hinges, a design and fabrication system for 3D-printed parts that sense their own deformation. A glove can turn finger movements into commands for a robotic hand, while a foldable lamp can use a pinch as a brightness control.[1][2]

MIT described the project on 22 October. The research, posted in September and accepted for the UIST 2026 conference, develops self-sensing compliant mechanisms: structures that bend through their material rather than relying on a conventional rigid hinge. X-Hinges makes their mechanical response and electrical sensing part of the same design problem.[1][2]

The hinge becomes the sensor

The method uses multi-material fused-deposition printing with two conductive filaments of different conductivity. Their arrangement inside the compliant structure creates electrical pathways whose resistance changes as the part deforms. Those changes provide information about bending and other motion, reducing the need to mount a separate sensor onto each moving element.[1][2]

Resistance is an intermediate signal. To turn it into a useful measurement, the system needs calibration that connects the electrical response to the motion of that particular design. The researchers combine data acquisition with regression models, allowing the sensing behaviour to be fitted to the printed structure. The paper describes continuous sensing across multiple degrees of freedom.[2]

Designing the response before fabrication

The software helps a designer choose the mechanism's geometry, mechanical properties and motion-sensing requirements. That is a useful shift for interactive objects: the question is no longer only where a sensor will fit after the object is made. The shape, movement and sensing can be developed together.[1][2]

The demonstrations explore different forms of that idea. Finger motion in a glove controls a robotic hand. An origami-style lamp responds to being pinched. Other prototypes use load or tactile signals to classify objects and translate the movement of a shark-tail controller into an input. Each turns a physical property of the object into an interface.[1]

These examples matter because ordinary objects often have very little room for separate electronics. Integrating the sensing material can make an interaction fit the object's existing form: squeeze the part that already flexes, or bend the element already designed to move. It also gives designers a common process for developing several kinds of movement-based control.[1]

The engineering question is repeatability

MIT identifies variation between prints as a calibration bottleneck. Two objects that look identical can have slightly different electrical layouts and readings. The team developed a computer-vision method to align the outputs and hopes to build recalibration into the objects themselves. Reproducible measurements matter if the same design is to work for different users or across multiple devices.[1]

The researchers also propose combining actuation and sensing in one print, so a structure could both move and report its movement. X-Hinges supplies the sensing part of that direction: a way to make the physical object's shape and deformation carry more of an interface's function.[1][2]

Report an error or suggest a correction