Today, knitting machines are used to produce many identical garments in bulk. They require knitting experts to design each pattern as well as engineers to look after the machinery.
“If you run a floor of knitting machines, you also have a department of engineers,” said Professor James McCann, who leads the Carnegie Mellon Textiles Lab. “It’s not a sustainable way of doing one-off customised pieces.”
McCann and his colleagues in the lab began work on software for generating instructions for V-bed knitting machines from 3D meshes: a standard way of modelling 3D shapes. V-bed knitting machines use hooked needles to knit yarn, which rests in needle beds angled in the shape of the letter V. These machines are highly efficient, but not very adaptable.
The team built their algorithm to take the machines’ limited capabilities into account, and to minimise the risk of yarn breaking or jamming.
So far, they have experimented with their system and produced knitted plush toys and clothing. Next, they hope to develop the algorithm further, such that it can work to knit cloth with the distinctive texture of knitted garments, rather than only using smooth, fine stitches.
According to McCann, the capability of the system to generate knitting patterns without human input could make on-demand machine knitting possible. He imagines that the knitting machines in factories today could instead be used to create customised, unique pieces in small batches such as perfectly fitting gloves.
According to McCann, front-end design systems such as this – which use common software and file formats to run equipment and produce unique designs – are common in 3D printing and machine shops, although they have not caught on in the world of knitting, in which languages and tools are not consistent. A push to develop knitting software such as this could allow machine knitting to be as adaptable as these other processes.
“Knitting machines could become as easy to use as 3D printers,” McCann suggested.