A vast Amazon warehouse on the outskirts of Tracy, Central California, is waiting to be handed over by a robotic arm. The robot then throws a Pulcel’s down chute to be loaded into the delivery truck and leaves.
“Each of them is their own beast,” says Jonathan Rammy, the robot he maintains. “They go through thousands of packages per hour and they don’t give up.”
These robots play a bigger role in warehouse networks, with artificial intelligence built into them, allowing large companies to expand these technologies in search of beating their convenient rivals.
Amazon is testing more sophisticated systems as part of its drive to improve the efficiency of sorting and satisfying centers. Investments in robotics and AI are part of the e-commerce giant’s $100 million planned capital expenditure this year.
It took Amazon over a decade to reach this point.
Amazon has expanded its robotics operations since it acquired Kiva Systems in 2012 © Cayce Clifford, FT
Why did Amazon start using robots in warehouses?
The Seattle-based group is currently one of the world’s largest industrial robot users, and its appetite for technology shows no signs of decline as it seeks to meet CEO Andy Jussy’s ambitions to significantly cut delivery times.
“Our robotics touches on the billions and billions of packages that we ship each year. This is a ‘great flywheel’ that increases productivity while reducing strain and monotony for human workers, he adds.
Amazon has dramatically expanded its robotics operations since it acquired startup Kiva Systems in 2012 for $775 million. Since rebranding AmazonRobotics, the division has deployed over 750,000 mobile devices and tens of thousands of robotic arms and autonomous systems.
The company has over six different robotic systems for logistical operations, ranging from heavy lift units and sorting machines to robotic arms.
What can Amazon robots do?
The first mobile robot deployed in Amazon’s warehouse was the Kiva robot, which moves by lifting the shelf unit from below. The system has evolved into what is called Hercules and Titan robots, some of which can be lifted with tons.
It also deploys a robotic arm, such as a cardinal, which pairs the suction cup with a set of sensors and cameras, determining the pressure needed to safely lift and move the item without causing damage. Other machines help you move, package and sort items as efficiently as possible.
Amazon’s robots
©Amazon
Amazon’s Sequoia system takes plastic wooden boxes containing stock from shelf and makes it quick to grab items to hand them over to employees
©Cayce Clifford, For the Ft
Small “drive units” can also be brought across the warehouse to designated chutes, and shelving units can be lifted and brought to employees.
©Cayce Clifford, For the Ft
Robotic arms such as Robin, Sparrow, Cardinal Select select and organize items for employees to box up or load into delivery trucks.
©Amazon
Amazon’s Proteus Drive unit (photo) is trained using artificial intelligence, deploying computer vision and allowing employees to freely navigate the space
The custom-made software and tools developed by Amazon are built on hardware manufactured by companies such as ABB in Switzerland and US-based Fanuc.
Robotics is implemented at each stage of Amazon’s fulfillment process, but the company claims that human workers continue to play a role in their operations.
What are the impacts of workers?
Eva Ponce, director of the Center for MIT Transportation & Logistics, says the technology is replacing several jobs, particularly more manual tasks, including heavy lifting, but also offers new opportunities within warehouse operations.
“Companies are investing more in their high-end peers. Make sure they’re ready for new styles of work,” adds Ponce. “More complex tasks still need to be done by people.”
Amazon workers must handle packages that are clumsy labeled and packages that have been processed with returned items that need to be inspected for damage.
The group says it has spent about $1.2 billion since training workers in 2019 for maintenance roles. It opposes the notion that it is trying to replace all workers with machines, noting that it employs more than a million people than when it purchased Kiva in 2012.
However, working conditions within warehouse operations are the source of constant tension with labor organizations, warning that robotic warehouses will increase the pace of work each day and push workers beyond the limits of what is safe.
A report from the US Senate Committee issued in December showed that Amazon warehouses recorded an injury rate of 30% higher than the industry average.
Amazon said that robotics and AI will “play an important role in supporting company-wide safety goals,” and that the technology hopes to improve safety through better ergonomics and heavier lifting.
Workers operate machines for packages that the automated system could not handle for FT.
How efficient is a warehouse robot?
Amazon’s investment in robotics is primarily motivated by a willingness to meet the ambition to deliver several items within hours, using increased efficiency to keep prices down, and to reduce costs.
Jassy told investors in a recent revenue call that the business will continue to invest in robotics and automation.
The warehouse, which opened last year by the company in Shreveport, Louisiana, has 10 times more robotic equipment than previous versions. It is estimated that the investment reduced the cost of fulfilling the order by 25%.
This type of warehouse operations will generate approximately $10 billion in annual savings by 2030.
What about this changing e-commerce?
Amazon’s rivals, including Walmart, have also invested heavily in robotics and automation, and Ponce notes that the desire to be more efficient and productive amidst the labor shortage is increasing.
“These technologies are disruptive,” Ponce says. “The warehouse of the future is a combination of robotics, sensors and computer vision.”
Brady says Amazon is using artificial intelligence to further advance. Its Proteus units use “computer vision” – helping to navigate spaces with human workers by deploying AI to allow robotic systems to process their field of vision in real time.
Proteus allowed the company to reduce the size of its warehouse, but by creating a “digital twin” (simulated version of the warehouse) alongside chip maker Nvidia, it was able to train autonomous robots at unprecedented speeds.
“This completely changed the way we do business… (and) it’s just the beginning,” Brady says.
