Robotics: Ready for warehouse prime time

Robots offer an appealing vision of the warehouse of the future. Repetitive, strenuous tasks like manual palletizing are instead predictably executed by a tireless robotic arm, freeing up humans to fill more value-added roles. Instead of a worker loading and unloading floor-stacked trucks in sweltering heat, a robot might reliably ferry product into and out of a facility around the clock. Even picking, where laborers work to fill a growing number of orders with smaller line counts, could be handled by a dexterous machine with more accuracy in a smaller footprint.

But as customers and robotics suppliers have learned, it is not as simple as installing a robot where a human once stood. The consistency and repetition of manufacturing processes, where robots have a strong foothold, are nowhere to be found in the dynamic environment of the warehouse. Palletizing, for instance, is not just a matter of picking things up and putting them down. A pallet must be built with an eye toward stability, preventing product damage and, increasingly, store-ready sequence.

Meeting the challenge in these applications requires three elements working in unison: the robot, its end-of-arm tooling (or end effector) and software. Recent advances in each of these areas are building the case for robotics in warehousing and distribution. “The technology is ready for prime time,” says Terry Zarnowski, sales and marketing director for Schneider Packaging. “I think we’re going to see it explode in warehousing in the next few years.”

Yet, to be truly successful, a robotic implementation must take into account the processes on either end of the robot itself. “A lot of times an end-user will call with a very specific process they want to automate,” says Alfredo Valadez, manager of business development and sales for the robotics division of Wynright. “With labor-intensive palletizing, the answer is yes, we can automate that, but we should be taking a wider view of the product flow. Some improvements upstream and downstream can improve palletizing’s effectiveness.”

Putting the pieces together
Robotic solutions are often viewed as an alternative to costly manual labor, assuming an adequate pool of qualified workers is even available. Faced with an aging workforce, many employers struggle to fill positions with a notoriously high turnover.

“Sometimes labor is a dollar issue and sometimes it’s just a matter of keeping a facility staffed,” says Larry Sweet, chief technology officer for Symbotic. “The growth in robotics is coming from the meeting point between pressures outside of a customer’s control and the more competitive cost and capability of robotics.”

For many users new to robotics, palletizing is a natural starting point. Manual palletizing is ergonomically challenging, with bending and twisting that can easily lead to injuries. To maximize a trailer load, an operator might lift cases overhead to reach an 88-inch to 100-inch pallet height, says John Schwan, director of sales and marketing for QComp Technologies. A robot can perform these functions faster and without concerns for injuries.

However, a skilled human is also solving a three-dimensional puzzle while palletizing, selecting cases based on weight and dimensions. To compete, a robotic palletizer must be fed by a storage and retrieval system that can deliver items in a very specific sequence—and fast enough to keep up with the robot, says Sweet. Without forethought, a robotic palletizer might build a pallet quickly, but in a way that creates more work downstream. Non-sequenced pallets can increase delivery times, fuel costs and the amount of labor required at the retail store.

“Palletizing automation is a mature technology that’s ready to go right now,” says Earl Wohlrab, palletizing and robotics systems product manager for Intelligrated. “The challenge has been feeding those palletizers from a distribution stream. You can’t treat each area as an island. You can’t deploy automation in one place and not consider the process that feeds it and the process that takes it away.”

Integrated robotic software helps to visualize the finished pallet in advance, as soon as the order is received. Software then communicates to the picking operation the sequence in which each case needs to be retrieved. Technologies such as high-end sortation and automated storage and retrieval systems (AS/RS) make this happen, and without them many operations might struggle to properly feed a palletizing robot.

That said, many customers using a warehouse management system (WMS) do already have helpful software capabilities in place. In fact, says Zarnowski, the software packaged with the palletizing robot can become the master and work with the WMS to drive order selection, even for traditional picking operations.

Finding the right tool for many jobs
Once cases arrive at a station, the robot needs the correct tooling to successfully handle each case. In applications such as food distribution, this can be challenging due to the extreme diversity of packages. The tooling may be set up to automatically configure itself, swapping heads from case handlers to tray handlers and even vacuum heads for bags. “Each time it makes that change it adds to cycle times,” says Zarnowski, “but again, the software can prevent frequent back and forth by presenting product in like groupings.”

Traditional picking processes perform better when working with layers, since a picker can retrieve a number of cases at each location instead of traveling between each case. The same is true for traditional robotic solutions. But the market for layer-picking solutions is rapidly dwindling amid a transition to mixed-case pallets, according to Brian Keiger, sales manager of general industry for Grenzebach Corp. and chairman of the order fulfillment solutions council for Material Handling Industry (MHI).

For example, instead of eight stock keeping units (SKUs) in each pallet, it’s now more common to see eight in each layer. The resulting uneven surface has led to a number of robotic solutions that place one case at a time, in the same way as a human would. New tooling enables the manipulation of six cases at a time, each of which can be adjusted to different heights and orientations, according to Keiger. At the end of the robotic arm, six clamping “fingers” can move up, down, in, out or sideways to simultaneously deposit cases on an uneven surface.

A palletizing robot might also be combined with other automation technologies for even greater benefits, says Schwan. One of his customers used a pallet dispenser to automatically supply pallets, while a short conveyor transported finished pallets from the robot to an automatic stretch wrapper. This eliminated several steps previously handled by forklifts.

Robotic selection
With software directing palletizing, a robotic system can create predictable cycle times. For instance, a certain case will always take precisely 6 seconds to place on a pallet. Manual picking upstream might struggle to match this clockwork precision. “Where it really gets interesting is when robotic picking and palletizing work together,” says Zarnowski.

With an average grocery order, a good selector with a double pallet jack and voice system will pick about 120 cases per hour, says Sweet. “If you have a facility moving a million cases per week, do the math and that’s a lot of direct labor.”

Early efforts are already underway to replace that selector’s traditional pallet jack with a robotic unit that allows a worker and equipment to work as a team. Taking it one step further, Wohlrab envisions a future where a robotic arm could be mounted to the pallet jack to pick and palletize in one step. “A few years ago that was science fiction, but now we talk very seriously about making that happen,” he says. “That’s ultimately where we’re headed.”

In the meantime, robotic pickers might take two forms. If stationary or rail-guided, a robot could replace a worker in a goods-to-person setup or a pick module. Alternatively, instead of waiting for product to be delivered and presented to the robot, it could be integrated into an AS/RS to retrieve cases individually. Such an approach is easier with case storage systems, but current machine vision technology can enable a robot to retrieve a case from a single-SKU pallet, says Zarnowski.

Although robotic pickers are currently well-suited to case handling, the industry is transitioning to more each handling, which is a bigger challenge for robotic vision and manipulation technologies. “Solutions are in development for purely robotic bin picking and they’re gradually coming onto the market,” says Keiger. “We’re seeing a huge uptick in the need for that, driven by e-commerce. People don’t order pallets online. They order pieces.”

Robotics for receiving
As another place with intensive case-by-case handling, the dock is also ripe for robotic automation. But while conventional palletizing brings the work to the robot, dock robots must go to the work. “Before, the operator had to walk 40 feet or more carrying each case to produce pallet loads,” says Valadez of one customer. “The customer now uses a palletizing robot on a linear track instead of a stationary mount, allowing the robot to build to multiple pallets simultaneously.”

Robots can also enter the trailer to retrieve pallets or floor-loaded cases and present them to a depalletizer or palletizer. Using advanced perception technology, the robot navigates into a trailer while extending conveyor behind it. The same vision allows its arm to retrieve one case at a time and place it on the inbound conveyor. From there, cases might be conveyed to reserve rack storage, a pick face, or directly to an outbound order. “With fully automated receiving, we’ve taken out all the driving around, the review of the manifest and the decision making as to where each item goes,” says Wohlrab.

From the inbound trailer to the outbound parcel carrier, robots appear in a growing number of fully automated facilities. Keiger notes an increase in the number of customers going to fully automated storage, retrieval and palletizing, with minimal human intervention. One robotics customer saw a 50% increase in picking rates in the first six months, he says, before ending up about 20 times faster than manual picking.

Such facilities are the rarity now, but Wohlrab predicts robotics will make inroads fast. “In another three to five years we should see robotic solutions installed in distribution centers that are true working platforms,” he says. “We’re probably five to 10 years from seeing robots as widespread in distribution as in manufacturing.”

Companies mentioned in this article
Grenzebach Corp.: grenzebach.com
Intelligrated: intelligrated.com
QComp Technologies: qcomptech.com Schneider Packaging Equipment: schneiderequip.com
Symbotic: symbotic.com
Wynright Corp.: wynright.com