Why Robots Can’t Sew Your T-Shirt
SoftWear Automation a robotics company that wants to make T-shirts. “We want to make a billion T-shirts a year in the US, all made as demanded,” said SoftWear CEO Palaniswamy Rajan.
The company was launched in 2012 with help from Georgia Tech Advanced Technology Development Center and a contract with Darpa. Two years later, a prototype is already in operation. Come 2017 work began on creating a production line capable of making shirts. That same year, the company negotiated with a Chinese garment manufacturer to set up a large production facility in Arkansas. That deal is over, though, and SoftWear is now focused on opening its own clothing factories.
The length of time it takes to reach this point is not surprising. The machines have proven to be proficient in many steps of making clothes, from printing on fabrics on cutting the fabric and folding and wrapping the finished garments.
Although sewing is very difficult to automate, because the fabrics bind and stretch their work. Human hands are skilled at keeping the fabric organized as it passes through a sewing machine. Robots often do not have enough power to control the task.
SoftWear robots have overcome obstacles. They can make a T-shirt. But making them as cheap as working people do in places like China or Guatemala, where workers earn as little as they can in the U.S., is a challenge, said Sheng Lu, a professor at study of fashion and clothing at the University of Knowledge
SoftWear calls the robotic systems Sewbots. These are usually augmented work desks that match sewing machines with complex sensors. The company diligently pays attention to the details of how it works, but here are the reasons: The fabric is cut into pieces that become part of the shirt: the front, the back, and the sleeves. The pieces are loaded in a line of work where, instead of someone pushing the fabric through a sewing machine, a complex vacuum system stretches and moves the material. The cameras track the insides of each panel, allowing the system to make repairs while the garment is being built.
But no two sets of cotton are exactly alike, always different from harvest to harvest; the variations in fabric and dye further complicate matters. Each variation may require the system to be reconfigured, disrupt operations, and SoftWear must train its machinery to respond accordingly. “The biggest challenge we face going into a production system is the need to operate 24/7 at high speed and above 98 per cent quality,” Rajan said.
Textile factories put out more than 20 billion T-shirts a year, mostly outside the US. In order to be able to make a T-shirt in the US, it has to be cheaper than importing. But eliminating the costs of shipping and import duties is not enough to cover the cost of paying U.S. workers to sew clothes. The Bureau of Labor Statistics says the average sewing machine operator in the United States makes $ 28,000 a year. About $ 13.50 an hour-more in countries where many T-shirts are made today. Lu, the professor in Delaware, said the Chinese salary for this type of work is almost one -third of the U.S. wage, while in Guatemala they are not one -fifth of the U.S. wage.
The focus on T-shirts allows SoftWear to avoid a common problem of automated sewing systems: switching from one type of garment to another. A skilled group of people can sew men’s shirts that are short a little one day and that woman’s the next. Such changes are especially challenging for robots. The way a cotton polo is sewn together is different from how a pair of polyester pants is made. Creating a new line of work for different fabric cuts and sewing different seams is complicated and expensive. Once production is ready to make T-shirts, it’s hard to easily configure Sewbots to make more.
Since the initial funding, SoftWear has raised $ 30 million in investments and investment grants – including a $ 2 million grant from the Walmart Foundation. Rajan said it would take ten million more to get the production of 1 billion T-shirts annually. To reach the target, the company needs several facilities, each with its own Sewbots and skilled workers to maintain it. Rajan says a Sewbot line of work can make a T-shirt every 50 seconds. At that rate, if it continues to run, a single line of work could produce more than 620,000 T-shirts per year-meaning it would take 1,607 Sewbots to keep working to reach 1 billion a years. Rajan said a more realistic number is close to 2,000; so far, the company has earned as little as 50.
Robots inevitably raise suspicions of evicting people and destroying jobs. Rajan acknowledges that SoftWear will employ fewer people than a traditional t-shirt, but he believes his company will create higher-paying jobs for people who keep the machines. “You want to improve the workers, and you want to train the workers,” he said. “Our goal is to have skilled workmanship and fast, agile production.”