Driverless forklifts are now a thing

In 2012, our article "Leave the (forklift) driving to us?" reported on the driverless forklifts that were then beginning to appear at trade shows, attracting both interest and skepticism. Eight years later, these machines have moved beyond the novelty stage and are increasingly being deployed in warehouses to supplement or replace human lift-truck operators.

It's unknown how many driverless (also known as automated or robotic) forklifts have been sold so far, but the list of companies that offer them is quickly growing. (For a partial list, see the sidebar "Who sells them?" in the online version of this article.) It's significant that some of the world's largest forklift OEMs, like Toyota, Crown, Jungheinrich, Kion, Hyster, Raymond, and Yale, are signing on; these companies are innovators, but they don't commit resources to a new product unless they're confident there will be significant, long-term demand. Other indicators that the outlook for this equipment is strong are acquisitions like Teradyne's recent purchase of the robotic forklift maker AutoGuide Mobile Robots.

Many observers believe that automated forklifts are on the verge of making significant headway in warehouse applications. So why are driverless industrial trucks getting so much attention now? Who is adopting and implementing them? And what do potential buyers need to know before they make the leap?

POISED FOR A BREAKTHROUGH?

For the purposes of this article, we're defining driverless/automated/robotic forklifts as industrial trucks that are equipped with forks and can do the following: transport, lift, lower, put away, and/or retrieve loads, either in fully automated mode, without a human operator, or in semi-automated mode, with a human operator working alongside them. These forklifts are a type of automated guided vehicle (AGV), but that term is more commonly used to describe conveyances that typically (but not always) lack forks and only provide horizontal transportation. We're avoiding the term "autonomous," which does not apply to much of the equipment discussed in this article. An autonomous vehicle operates more independently, "thinking" and making decisions, notes Tiger Xu, product manager, automated solutions, for The Raymond Corp.

Interest in automated forklifts is growing, first and foremost, because the U.S. labor market is so tight. Unemployment is at its lowest point in 50 years, and warehouses and DCs are struggling to hire and retain reliable, qualified employees. Automation can help fill that labor gap, providing continuous, consistent performance 24/7. It can also mitigate the cost of finding and onboarding new workers. A bonus: Adding capacity doesn't require adding operators. "Some of the key advantages of automated trucks are their flexibility, scalability, and reliability," Xu says. "They are easily scalable up or down; if you want to add another 20 trucks, you simply code in the data and they will quickly be able to do the same tasks in the same way as the original trucks," he says.

Second, automation can reduce product damage and alleviate the ergonomic impact of repetitive, physically demanding tasks in today's fast-paced warehouses. And third, the enabling technology has made great strides recently. Today, automated forklifts can locate specified storage slots and retrieve or put away loads with pinpoint accuracy, stop and alert humans when they encounter obstacles or anything else unexpected, choose the most efficient travel path, and remember where empty slots are located, among other capabilities.

"Supervisor" software controls task allocation and the robotic forklift fleet's movements. The software typically interacts with a warehouse management system (WMS) and may also interact with enterprise resource planning (ERP) systems and with other material handling equipment. Some forklift OEMs have developed their own navigation and control systems, while others partner with specialized tech companies. Examples of the latter include the recent alliance of Crown Equipment Corp. and JBT Corp., and partnerships between The Raymond Corp. and Seegrid, between Balyo and Hyster-Yale Group, and between JBT and Yale. Such collaborations integrate both partners' domain-specific expertise and leverage different approaches to problem solving, allowing innovative products to come to market more quickly than they otherwise could, says John Rosenberger, director of iWarehouse Gateway and global telematics at Raymond.

A key factor in gaining wider acceptance is technology that allows robotic lift trucks to operate without following fixed paths defined by guide wires, magnetic tape, magnets, or transponders embedded in the floor. Newer guidance technologies include methods such as lasers, image recognition, and location sensors that measure distances, map facilities, store data, and navigate based on the forklift's position relative to "natural features"—walls, racks, and other objects. Depending on the technology, minimal or no additional infrastructure is required, so it's easy to reprogram travel routes.

Of the various options available today, LIDAR (light detection and ranging), which uses lasers to measure distances to an object, seems to be gaining the most traction in warehouses and DCs. "There are others, like vision-guided systems, but nothing seems to be taking hold as well as LIDAR, because of its accuracy and low cost of implementation," says Lou Micheletto, sales manager-robotics and automation for Hyster Co.

In addition, the wider availability and declining costs of sophisticated sensors have reduced costs for the end-user. "Before, only three-shift or four-shift operations with very high hourly pay rates could find justification for an automated forklift project," says Mick McCormick, director, robotics and automation, for Yale Materials Handling Corp. Now, a wider range of operations, including those with just a few trucks, can get a return on their investment, he and other advocates say.

Not everyone is jumping in yet. The third-party logistics service provider Kenco, for example, has evaluated automated forklifts but is not using them in any of its DCs, says Kristi Montgomery, vice president, innovation and research and development. They do hold promise, she says, noting that advances in sensors and the Internet of Things have improved the capabilities of automated vehicles of all kinds. Furthermore, insights gleaned from the new data generated by driverless forklifts will improve decision-making. "But there are still concerns around safety and the capabilities provided for the cost invested," she says. "The technology is still new, and so there are still things to be worked out from a usability standpoint."

WHERE DO THEY SHINE?

Currently, fully or semi-automated counterbalanced forklifts, narrow-aisle forklifts, high-reach trucks, and high- and low-platform pallet trucks are on the market. While automated forklifts can be programmed to do many tasks—Linde says its "Matic" product line covers "80% of all handling tasks in production and storage facilities"—they aren't suitable for every environment or job. OEMs have zeroed in on activities where driverless trucks can improve productivity and accuracy, reduce injuries and damage, and handle repetitive, non-value-adding tasks. These include:  

• Horizontal transport, with or without loads, between specified points, with no other action required
• Pallet picking and putaway
• Case picking and putaway (typically using semi-automated trucks that users control remotely as they pick orders).

Some end-users are focusing on horizontal transport, says Micheletto of Hyster. Examples he's seen include delivering pallets to manufacturing areas or storage, moving loads from storage to shipping, and transporting loads from conveyors and automated storage and retrieval (AS/RS) systems to a stretch-wrap station.

According to Raymond's Rosenberger, there's more demand for automation in receiving and replenishment, which largely involve routine processes and thus are well suited to automation, than there is in picking, a high-speed activity that requires flexibility. He's also seeing adoption at combination manufacturing/warehouse facilities, where product comes off a production line in batches and moves into an attached warehouse.

Consumer packaged goods and food and beverage manufacturing and distribution operations have been the biggest users of Yale's robotic forklifts, McCormick says. He's seeing pilots where a DC will conduct a limited test and then look at scaling up, while a few customer locations have done "full-scale, complete, almost 'lights out' implementations," he says. Those operations continue to manually load and unload trucks, but robotic forklifts handle everything inside the DC. A small staff supports the robotics system and resolves any problems. Facilities that are running completely robotic, he notes, tend to work 24/5 or 24/7, with more than two shifts.

POINTS TO PONDER

Thinking about giving driverless forklifts a try? It's a big leap for most operations, with many complex considerations to keep in mind. Here are some of the big ones:

• Prioritizing the steps. One piece of advice that came up over and over was this: "Don't just automate your current process. Optimize first, then automate." In Rosenberger's view, this may be the single most important bit of advice. He strongly urges those planning to introduce driverless forklifts to "connect, optimize, and only then automate—in that order." By "connect," he means using telematics and fleet management systems to get accurate baseline data about forklift activity, which can then be used to improve current operations. That step will help users make significant headway on the problems they want automation to solve, freeing up money to spend on the project while helping identify where automation is genuinely needed and can generate an ROI (return on investment), he says.
• Cost justification. ROIs of two years or even less are achievable, according to some proponents. But other observers caution that overall costs can be very high, because end-users must consider the costs of the entire "package," not just the new equipment. If automation makes sense but cost is a concern, using partially automated equipment or implementing in a limited area where automation will have the greatest benefit are options.
• Operating speed and efficiency. According to one OEM executive, who spoke off the record, automated forklifts are too slow and regimented for the fast pace and frequent pivots necessary in many DCs today. There's also potential for traffic congestion if the travel speeds, number of trucks, and traffic patterns aren't calibrated perfectly. Furthermore, the ANSI/ITDSF B56.5 safety standard for AGVs applies to robotic forklifts. That means they require a larger "envelope" of operating clearance than human-operated forklifts do, says another executive, who also spoke off the record. In a dense operating environment, smaller forklifts may be needed to maintain the minimum required clearance, he says.
• Equipment adaptability. Kenco's Montgomery recommends thinking about automated forklifts' "conversion capability"—"can they be used as traditional lifts in difficult scenarios when a human needs to intervene or are they autonomous only?" (Most manufacturers say their automated forklifts can easily be switched to manual mode.) She also suggests considering whether the forklifts can adapt to dynamic physical environments and product profiles.
• Human oversight. A team that includes an experienced project manager from the end-user, the forklift dealer, and the systems integrator is a must. On top of that, at least one "tender" or robot supervisor is necessary in every robotic forklift installation, says Micheletto. This person takes care of exceptions—if there's a problem with the navigation system, for example, or a truck stops operating because it has encountered an obstacle (shrink wrap covering a fork slot, for instance).
• Maintenance and technical support. In addition to mechanical maintenance and repair, software upgrades and updates will be required. The forklift dealer should be capable of providing most of the support and service, but in some cases, the navigation-technology provider or the systems integrator may be involved. Do all of them offer local support?
• Readiness to work with robots. Everyone in a facility—"even the receptionist and accountant"—should be trained in how to interact with robots, says Micheletto. The "rules of the road" for automated forklifts are different from the guidelines for traditional vehicles; for safety's sake, they must be clearly defined. Even though automated forklifts are equipped with sensors to prevent collisions, human beings are unpredictable, so enforcing the rules is critical. In an interview with DC Velocity, Melonee Wise, CEO of Fetch Robotics, recommended that training in the dos and don'ts of interacting with the robots include an explanation of what the machines can and cannot do (they can't "see" glass, for instance).
• A true picture of forklift operations. An accurate baseline picture of forklift operations is a must-have before robotic forklifts take over from human operators, McCormick says. In his experience, though, most facilities have gaps in their understanding of exactly what their forklift operators do. As a result, some things operators do aren't documented or included in a process map. For example, in addition to the obvious jobs like unloading pallets from a truck and putting them into storage, operators might also be labeling pallets or latching trailers to docks. Because a robotic forklift can't do those things, McCormick says, you'll have to find a person to handle those tasks.

MORE THAN A MACHINE

An effective automated forklift project requires recognizing that these installations are not simply about the equipment. Rather, they should be approached as complex solutions encompassing equipment, control software, wireless communications, sensors, specialized navigation technology, systems integration and technical support, and sometimes layout reconfiguration.

In the end, every automated forklift project will be unique and complex. Careful preparation, accurate information, and a realistic, "eyes wide open" understanding of what automation can and cannot do are important success factors. Or, as Yale's McCormick succinctly puts it: "Don't make assumptions."