Warehouse 4.0: Smart tech brings transparency to global supply chain

Making the right pick

For anyone seeking information about today’s warehouse automation, Amazon would be a logical first stop. Recognizing early on that it would need a boost in automation to meet skyrocketing e-commerce demand, the Seattle-based firm purchased Kiva Systems, a robotics company from Boston, for $775 million in 2012.

Kiva builds deceptively simple, but highly advanced, robots for reducing the movement of human workers inside Amazon’s enormous distribution centers. The individual Kiva unit – a squat, one-foot-tall orange rectangle – is not much to look at, but it can fit under a standard Amazon shelving unit, lift it a few inches off the ground and move it virtually anywhere within the fulfillment center, following the bar-coded paths embedded in the floor. Basically, instead of having thousands of human pickers run around these vast warehouses spaces to find the right bins for each order, Amazon uses the battery-powered Kiva bots to bring the stacks to the pickers, as needed, thus reducing worker fatigue, lowering the risk of injury and improving accuracy. Currently, the online retail giant operates any army in excess of 45,000 Kiva robots, working individually or in concert, at Amazon’s hundreds of fulfillment centers worldwide.

(For a glimpse at the Kiva ‘bot ballet that occurs at every Amazon fulfillment center, check out this mesmerizing video below)

One thing the super-efficient Kiva bots cannot do, however, is pick actual items off shelves. At the time the Kiva’s were built, no robot was yet smart enough to match the speed and accuracy of humans in pick-and-pack operations. But robotics in the warehouse business is rapidly changing, and IAM Robotics’ Swift unit is leading the way.

The Pittsburgh-based company completed its first commercial deployment in January – a months-long test project for the Rochester Drug Cooperative’s (RDC) distribution center in New York state, which supplies drug store items to various “mom and pop” stores across New England. During this trial, the Swift unit performed picks during the night shift at the RDC center, based on customer requests relayed from the facility’s warehouse management system (WMS) via IAM’s Swiftlink back-end interface.

Swift weighs about 180 kilograms, travels at about a meter per second and can make roughly 200 picks per hour. Using machine-learning techniques, the more picks Swift makes, the more efficient the routes through the warehouse become, said John Cameron, one of the head engineers at IAM Robotics. In a few years, he added, there will be a fleet of five Swift robots, all making picks at the same time and coordinating with the human pickers as well.

Competitors like InVia, Fetch, Locus and others are trying to achieve full autonomy, Wells said. However, Swift is the only robotic picker that is currently capable of doing unassisted work 24/7. “It’s the robot itself that’s doing the picking,” he said. “It’s not a follow-along solution, where there’s a human that still needs to be in the loop.”

During the RDC trial, the robot performed materials handling tasks at or above human speeds, and, of course, never needed sleep, got injured or got sick. On a dollars-and-cents basis, Swift has cheaper operational costs than an actual human, considering the healthcare, taxes, workers’ comp and other costs associated with warm-blooded employees.

For shippers and 3PLs that may blanch at the thought of adding a disruptive fleet of robots to a fully functional warehouse setting, Cameron says Swift is designed to work in current facilities without the need of a retrofit and can be up and running in just a couple of days.

Wells is not shy about expounding on the advantages of the Swift system (see photo above). “I challenge anybody in the industry to come and try to meet our picking rate in RDC,” he boasted. “I think it would be hilarious to have everyone else try to come and see who really is in front here.”