【AMNews】Researchers from MIT’s Computer Science and Artificial Intelligence Laboratory (CSAIL) and China’s Tianjin University (TJU) have launched X-Hinges, a design software that automates sensor integration for 3D-printed objects, solving a longstanding signal interference problem to enable multi-axis motion sensing in everyday items.
A core challenge in 3D-printed sensing devices has been material uniformity: sensors and their connecting traces are typically printed with the same conductive material, producing overlapping signals that make it hard for devices to distinguish actual motion from background noise. This flaw has long limited the accuracy of interactive printed products, from wearable gloves to robotic components.
X-Hinges addresses this by splitting the electrical workload across two separate filaments. One is optimized for sensors to deliver the strongest possible sensing signal, while the other handles trace routing without causing signal interference. Once users finalize their structural design, the software automatically places both materials in appropriate positions, with no manual wiring layout required.
The software follows an accessible drag-and-build workflow. Users assemble objects by selecting from four base shapes — cylinder, rectangle, thin outline, or custom form — and define how each movable section operates. Each hinge supports up to three axes of motion: compression, vertical bending, and lateral bending. The system then embeds sensors directly at the bending joints, capturing motion data from all directions simultaneously.
“Most existing methods only support one axis of motion, largely due to signal interference issues,” said Xiang Chang, lead author of the study and a TJU PhD student conducting research at CSAIL. “X-Hinges lets your 3D-printed device move and sense across three angles, opening up far more interactive possibilities.”
The research team has built multiple working prototypes to demonstrate the technology. A 3D-printed data glove adapts to different hand sizes, records finger joint bending in real time, and transmits motion data to a humanoid robotic hand to mimic grasping movements — a function that could speed up robot skill learning in factory and home scenarios. An origami-inspired lamp, which resembles folded paper when closed, adjusts its brightness when pinched by the user. The team also developed a tactile sensing platform that can identify objects placed on its surface, such as a banana, purely from pressure data, plus a cartoon shark-shaped game controller with its tail functioning as a joystick.
“It’s kind of like adding a conscious element to everyday objects,” said Jiaji Li, senior author of the paper, MIT EECS postdoc and CSAIL researcher. “What you get is objects that better understand how they’re being used and have more ways to respond.”
The technology still faces a calibration bottleneck. Slight variations in conductive element layout occur from print to print, meaning two visually identical parts may produce inconsistent sensor readings. Chang and Li have built a computer vision system to recalibrate outputs and align data across matching parts, with a long-term goal of integrating a self-calibration mechanism directly into the printed devices.
The work builds on Li’s prior research on 3D-printed tendon-driven actuation mechanisms. His next objective is to combine actuation and sensing in a single 3D printing process, creating a unified fabrication platform for embodied intelligence and physical AI agents.
Co-authors of the paper include Stefanie Mueller, MIT associate professor of EECS and CSAIL principal investigator, and recent TJU graduate Haiyang Yan. The team will present the findings at the ACM Symposium on User Interface Software and Technology (UIST) in early November.
