Constructing devices that mimic the fluid, accurate movements of humans is challenging, yet this could transform if we could manipulate a few (literal) strings.

At least, that’s the concept behind “cable-driven” systems, where threading a string through an object enables smooth motion across various sections of that object. Consider a robotic finger, for instance: A cable could be threaded from the palm to the fingertip and pulled to generate a curling action.

Although cable-driven systems can produce real-time movement to allow an object to bend, twist, or fold, they can be intricate and labor-intensive to assemble manually. To streamline this process, researchers at MIT’s Computer Science and Artificial Intelligence Laboratory (CSAIL) have created a comprehensive 3D printing technique known as “Xstrings.” Part design application, part manufacturing process, Xstrings integrates all components to create a cable-driven device, significantly reducing time spent assembling bionic robots, designing art installations, or developing interactive fashion concepts.

In a document to be unveiled at the 2025 Conference on Human Factors in Computing Systems (CHI2025), the researchers utilized Xstrings to produce a variety of vibrant and distinctive objects, including a red walking lizard robot, a purple wall art piece that can open and close like a peacock’s tail, a white tentacle that curls around objects, and a white claw capable of forming a fist to grasp items.

To craft these striking systems, Xstrings enables users to completely tailor their designs within a software platform, which they can then send to a multi-material 3D printer to realize their creation. Users can automatically print all components of the device in their preferred locations in one go, including the cables interwoven throughout and the joints that allow for desired movements.

MIT CSAIL postdoctoral fellow and primary author Jiaji Li states that Xstrings can conserve time and effort for engineers, cutting total production time by 40 percent as opposed to manual methods. “Our novel approach can assist anyone in designing and crafting cable-driven products using a desktop bi-material 3D printer,” claims Li.

A novel approach to cable-driven manufacturing

To utilize the Xstrings application, users initially input a design with specified dimensions, such as a rectangular cube partitioned into smaller sections with a hole at the center of each. Next, they can select the movement direction of these parts by choosing different “primitives:” bending, coiling (resembling a spring), twisting (similar to a screw), or compressing — along with the angle of these actions.

For more intricate designs, users can blend multiple primitives to generate captivating motion combinations. For instance, to create a toy snake, you might incorporate several twists to form a “series” combination, where a single cord dictates a sequence of movements. To develop the robot claw, the team embedded various cables in a “parallel” configuration, allowing each finger to close into a fist.

Beyond refining the movement of cable-driven mechanisms, Xstrings also streamlines the integration of cables into the object. Users can pinpoint exactly how strings are secured, selecting the locations for the “anchor” (endpoint), “threaded areas” (or holes within the structure for the cord to pass), and “exposed point” (where one would pull to operate the device). For a robotic finger, for example, the anchor could be positioned at the fingertip, with a cable running through the finger and a pull tag visible at the opposite end.

Xstrings further accommodates diverse joint configurations by automatically placing components that might be elastic, compliant, or mechanical. This adaptability ensures the cable can pivot as required to fulfill the device’s intended motion.

Driving exceptional designs in robotics, art, and beyond

Once users have simulated their digital schematic for a cable-driven object, they can proceed to fabricate it. Xstrings can direct the design to a fused deposition modeling 3D printer, where plastic is heated and extruded through a nozzle, allowing layers to be constructed gradually.

Xstrings employs this method to arrange cables horizontally and build around them. To validate their technique for successfully printing cable-driven systems, the researchers conducted thorough tests on their materials and printing conditions.

For instance, the researchers determined that their strings only failed after undergoing more than 60,000 pulls by a mechanical apparatus. In another experiment, the team found that printing at 260 degrees Celsius at a pace of 10-20 millimeters per second was optimal for producing their many imaginative creations.

“The Xstrings software brings a plethora of ideas to fruition,” remarks Li. “It allows for the creation of a bionic robotic device akin to a human hand, mimicking our gripping abilities. Additionally, you can craft interactive art installations, such as a cable-driven sculpture featuring unique geometrical forms, and garments with adjustable flaps. In the future, this technology could facilitate the quick, one-step construction of cable-operated robots in space, even in constrained environments like space stations or alien outposts.”

The team’s methodology provides substantial flexibility and a significant acceleration in producing cable-driven artifacts. It results in objects that are rigid externally yet soft and adaptable internally; in the future, they may explore the development of items that are soft on the outside but solid on the inside, akin to human skin and bones. They are also contemplating the use of more durable cables, and rather than solely placing strings horizontally, embedding them in angled or even vertical orientations.

Li co-authored the paper with Zhejiang University master’s student Shuyue Feng; Tsinghua University master’s student Yujia Liu; Zhejiang University assistant professor and former MIT Media Lab visiting researcher Guanyun Wang; along with three CSAIL members: Maxine Perroni-Scharf, an MIT PhD student specializing in electrical engineering and computer science; Emily Guan, a visiting researcher; and senior author Stefanie Mueller, the TIBCO Career Development Associate Professor in the MIT departments of Electrical Engineering and Computer Science and Mechanical Engineering, and leader of the HCI Engineering Group.

This study was partially funded by a postdoctoral research fellowship from Zhejiang University and the MIT-GIST Program.