It turns out a knitting machine can work a bit like a 3-D printer.
Researchers at Carnegie Mellon University's Textiles Lab have created an algorithm that translates 3-D shapes in computer models into instructions for automated V-bed machines, a common piece of industrial knitting equipment.
The algorithm, created by CMU Robotics Institute faculty member Jim McCann and Ph.D. students Lea Albaugh and Vidya Narayanan, produces stitch-by-stitch instructions for the V-beds.
When the piece comes off the machine, it’s flattened, with an opening on one end, almost like a sock. From there, the user can manually stuff the fabric with cotton to make it take on the 3-D shape from the original model.
"I want to live in a world where people can create custom soft objects like clothing or stuffed animals. And this is one step towards that world," said McCann. "Right now ... if you’re an expert you can design things by hand, but it takes quite a while and a lot of skill.”
Being able to design knit fabrics via computer software could make the process more accessible to non-experts and easier even for experts.
"[We are hoping to] minimize the amount of things that a user typically doesn't want to do — for example, thinking about stitch-level details— but also allow a person to do what they do want ... which may be adding more seams or have more creative control over the process," said Narayanan.
One challenge that still remains is being able to match the variation of textures that can be achieved by expert level knitters: currently the algorithm can only design for smooth knit surfaces.
"When we show this to people who are used to knitting, like hand-knitters ... they say, 'Oh, but it's just stockinette, which is the most boring kind of knitting,'" said Albaugh. "In the world of 3-D printing, we're like 'yeah, it's a smooth surface, that's what you've got.' But knitting inherently has a lot of other ways that you can manipulate a knit fabric that are cool and interesting and that we see in most clothing items.”
The three researchers said that in the future, this kind of technology could potentially be used to help people design clothes with a fit they feel comfortable in.