Robot tailors are on their way
IN 1970 William J. Bank, president of the Blue Jeans Corporation, predicted that there would be a man on Mars before the production of apparel was automated. Almost half a century later, he has not yet been proved wrong. Viewed through the lens of history, this is astonishing. Spinning was one of the first processes to succumb to industrialisation. Weaving followed shortly afterwards. Cutting the resultant cloth into pieces from which an item is then assembled is easy now that patterns can be reduced to software. But, though effective sewing machines have been around since the 1840s, their activities still have to be guided by hand. The idea of putting a bolt of fabric into one end of an automated production line and getting completed garments out of the other thus remains as impossible as it was in Bank’s day. Two American companies, however, think that they have cracked the problem, and that a system which can turn cloth into clothing without the need for tailors is just around the corner.
One of these aspiring firms, SoftWear Automation in Atlanta, Georgia, already makes machines, Sewbots, that can turn out towels, pillows, rugs, mats and other such essentially rectangular goods. SoftWear’s boss, Palaniswamy Rajan, thinks Sewbots are almost ready to take the plunge with actual garments—in particular, with T-shirts. The other aspirant, Sewbo, which is based in Seattle, has made a T-shirt already, as a proof of principle, though it does not yet have a commercial production system.
SoftWear’s Sewbots rely on two things: high-speed, high-resolution cameras able to monitor the movement of individual threads in a piece of cloth, and software that takes those movements and generalises them to describe the distortion and orientation of the fabric which the threads in question are part of. That permits a Sewbot to adjust the fabric appropriately as it is fed to the machine’s needles. This feeding is done by a vacuum-powered robotic grip, which can be programmed to work with a variety of fabric sizes.
Sewbo’s approach is to simplify the feeding process by stiffening the fabric. This is done by coating it with a thin layer of a plastic called polyvinyl alcohol before it gets anywhere near the needles. From then on, according to Jonathan Zornow, Sewbo’s founder, it can be handled as sheets of metal might be in a car plant. Aligning such sheets so that they can be sewn together is easy. And, once a garment is completed, the plastic can be removed by soaking in warm water. This is a routine procedure, for polyvinyl alcohol is already used elsewhere in textile production to strengthen yarn during weaving.
That both Mr Rajan and Mr Zornow have their eyes on the T-shirt market is no coincidence. T-shirts are structurally simple and demand for them is huge. Both of these things make them ideal for automated mass production. Mr Rajan thinks SoftWear’s Sewbot T-shirt assembly lines will be able to turn out 3,300 items a day. He hopes to start shipping them within two years. Twenty-one of them will go to a factory in Arkansas that belongs to Tianyuan Garments—a firm which is the largest producer of apparel for Adidas, a German sportswear firm.
Sewbo’s plans—those that have been publicly announced, at least—are less specific. The company’s demonstration T-shirt was made last year. Now, Mr Zornow says, they are refining their methods for various types of fabrics and industrial settings. Though the first use of fully automated tailoring will be for mass production, he foresees the technology becoming more specialised quite quickly. Removing the need for cheap human hands to assemble garments will permit factories to set up shop near customers in rich countries. That will allow rapid delivery to shops, and let retailers try things out in the knowledge that they can resupply successful lines almost instantly. In the longer run, the automation of garment-making may also usher in an era of “mass bespokeness”, in which customers choose a style and have it made to fit their bodies in a way only haute couture or Savile Row can manage today. Whether that will happen before a human being walks on Mars remains to be seen.
