Knitting Robot: From Stocking Frames to ABB Arms

Olympic diver Tom Daley told the Guardian in September that he is obsessed with his knitting machine, a Silver Reed LK150 he uses to make jumpers. Home knitters can buy a machine like his for a few hundred pounds, and factories have been knitting by machine for more than 400 years, so the idea of a knitting robot is less far fetched than it sounds. This post follows the story from hand knitting and William Lee's stocking frame to home knitting machines and Shima Seiki's WholeGarment machines. It also looks at where ABB robots such as GoFa, YuMi and Flexley Mover could tend the knitting floor, and what is still missing.

A computerised flat knitting machine at work in a knitwear factory, with yellow knitted panels hanging below. Photo: Asivechowdhury / Wikimedia Commons (CC BY-SA 4.0)

How knitters went from needles to frames

For most of history, every sock and stocking was knitted by hand on two or more needles. It was slow work done at home, often by whole families, and in places like Shetland it was a source of income. Hand knitting is still popular, partly because you can take it anywhere and change the pattern as you go. Daley himself made knitting famous again during the Olympics and later hosted a Channel 4 knitting competition, Game of Wool.

A daguerreotype of an elderly woman knitting with needles, from the mid 19th century. Photo: George Eastman House via Wikimedia Commons (no known copyright restrictions)

William Lee, who studied at Cambridge in the 1580s, built the stocking frame in 1589. It used a row of hooked bearded needles to form a whole row of loops at once, while a hand knitter makes one stitch at a time. Lee later moved to France with his brother and nine workmen, and Henry IV gave him a patent to make stockings in Rouen. After his death his workers came back to England, and framework knitting grew up around Nottingham and London.

A preserved hand operated stocking frame, with its long row of needles, in a museum display. Photo: ClemRutter / Wikimedia Commons (CC BY-SA 4.0)

Late 20th century domestic machines from brands such as Brother and Silver Reed held 120 to 200 latch needles in a flat bed, and a carriage pushed by hand knitted a full row in one pass. Punch cards, and later electronics, chose which needles knitted to make patterns. In factories, Masahiro Shima founded Shima Seiki in Wakayama, Japan, in 1962 to build fully automatic glove knitting machines. The company moved on to flat knitting machines and computer design, and in 1995 launched WholeGarment, which knits a complete garment in one piece. In 2016 Fast Retailing, the owner of Uniqlo, signed a deal with Shima Seiki.

A Brother KX395 home knitting machine with its carriage, tools and pattern book on a table. Photo: KS_aus_F / Wikimedia Commons (GFDL 1.2)

Knitting machines and robots on the floor today

Knitting machines are already automated, but people still feed and empty them. Knitting Industry reported on October 7 that the company will show its N.SSR112 computerised flat knitting machine and its SDS-ONE APEX4 3D design system at IGATEX Pakistan in Lahore from October 15 to 18. Its software creates virtual samples, and once a design is approved, the knitting data is converted straight into machine data. A second event in Milan, from October 14 to 28, shows yarns and digital knit samples.

A computerised flat knitting machine with its touch screen, in a factory full of identical machines. Photo: Asivechowdhury / Wikimedia Commons (CC BY-SA 4.0)

Robots fit best in the jobs around those machines. Yarn cones have to be loaded on the creel, finished panels or garments have to be taken off, checked, steamed, folded and packed. In a factory with dozens of machines, that is a lot of walking and lifting. Shima's KnitManager software already tracks each machine's status and production, so a factory can see which machine needs yarn or has a finished batch. That information is what a mobile robot needs to know where to go next.

Yarn cones feeding rows of flat knitting machines, the job robots could take over. Photo: Asivechowdhury / Wikimedia Commons (CC BY-SA 4.0)

ABB has not announced a knitting machine project that we found, so this is a possibility. A Flexley Mover AMR could carry yarn cones and finished garments between machines and the finishing room, finding its way with 3D visual SLAM. A GoFa CRB 15000 cobot on a cart could load cones or lift knitted pieces, working next to people because it stops when it senses contact. A YuMi could handle small, soft items such as socks or gloves. Our post on textile mill yarn handling covers the spinning side of this.

ABB's Flexley Mover P604 moves loads around a plant without fixed tracks. Photo: ABB

Where knitting robots go next

In factories, the next step is the dark knitting floor, where machines run through the night and robots refill yarn and clear output. WholeGarment helps, because it removes the linking and sewing steps that are the hardest for robots to handle. The market for the machines is changing too: Stoll, the German maker known for the first Nike Flyknit shoe, stopped making machines in October 2025. Soft material is the hard part: knitted fabric stretches, curls and snags, so a robot needs gentle grippers and vision to pick it up without damage. The same limit is why folding and packing still use so many people.

An ABB GoFa CRB 15000 cobot at a work table, an arm built to work next to people. Photo: ABB

At home, knitting robots are less likely to look like arms. Tom Daley's LK150 is a manual machine, and that is how most home knitters like it. A home robot could perhaps one day push the carriage or fix dropped stitches, but there is little sign of that yet. Research is going a different way. In September, scientists at Harvard described knitting that turns into shape changing smart fabric, and a Rice University engineer is exploring fabric that itself works as a robot.

A Knittax home knitting machine with two carriages and a project in progress. Photo: Gudde Fog / Wikimedia Commons (CC BY-SA 4.0)

Computerised flat knitting machines are expensive, and adding robots only pays off in large factories with many machines. Programming a knitting machine is a skill in its own right, and robot cells add another layer, so factories need technicians who understand both. Robot paths for loading and unloading can be tested in RobotStudio before they reach the floor. As for hand knitters, I doubt any robot will put them off, because for most of them the point is doing it themselves.

Programming a single arm YuMi with Wizard on the FlexPendant. Photo: ABB

Innovation

AI solutions for effortless ABB robot control.

Automation

Robotics

ceojohntran@chatwithrobot.net

+84905311611

© 2025. All rights reserved.

qtran1215@gmail.com