Meal kit market leader HelloFresh is speeding fulfillment, enhancing quality, and boosting menu variety thanks to an innovative automated system at its Irving, Texas, distribution center.
Victoria Kickham started her career as a newspaper reporter in the Boston area before moving into B2B journalism. She has covered manufacturing, distribution and supply chain issues for a variety of publications in the industrial and electronics sectors, and now writes about everything from forklift batteries to omnichannel business trends for DC Velocity.
Meal kit producer HelloFresh relies on automation to guarantee fresh and accurate shipments to customers—and that reliance has only increased as the company has grown from a small German startup to a global enterprise serving consumers in 18 countries. Surging demand, expanding menus, and the ever-present challenge of meeting high food quality and safety standards add complexity to the HelloFresh model, necessitating a focus on technologies that can give the business an edge as it grows.
Zeroing in on the U.S. market, company leaders took a leap nearly five years ago that would help HelloFresh meet burgeoning local demand and set the stage for further expansion of its menu and capabilities. They added a brand-new distribution center (DC) in Irving, Texas, that would feature the most advanced technology in the company's North American fulfillment network to date.
"Once we had identified that we were moving toward a greenfield site, we saw the opportunity to have an enhanced fulfillment system in the building," Kyle DeGroot, vice president of operations engineering and technology for HelloFresh, says of the Irving project. "We wanted the capability to expand our product offering while maintaining or increasing efficiency from a fulfillment standpoint."
The answer to that challenge: an automated storage and retrieval system (AS/RS) from AutoStore. The system is now the centerpiece of a customized, high-tech fulfillment process that is speeding operations, increasing throughput, and improving productivity—all while giving HelloFresh the flexibility to shift and expand its menu without complicating the fulfillment process.
TAMING COMPLEXITY
HelloFresh was founded in 2011 in Berlin and has grown from its early days as a community-based business into a global organization that delivered more than 1 billion meals to customers in 2023, according to company data. Faced with escalating demand in the U.S., HelloFresh set out to expand its fulfillment network in 2020, adding the 377,000-square-foot Irving facility and bringing the company's U.S. fulfillment network to a total of eight DCs.
More than that, the Irving DC was an opportunity to advance the company's use of automated warehouse technology and build on its presence in the Dallas-Fort Worth metropolitan area—both because of the facility's size and because it was a greenfield site, offering a blank slate for innovation. The goal was to leverage the reliability, density, and speed of AutoStore as the heart of an automated meal kit assembly process, according to HelloFresh and leaders at Swisslog, the material handling automation specialist that designed and installed the AutoStore system.
For Swisslog, the project represented a departure from typical AS/RS installations, which are designed to handle traditional distribution center operations, including e-commerce and store fulfillment.
"HelloFresh is in the distribution business, shipping meals to the home. However, once we took a closer look at [the business], we [realized] that it acted more like a production operation," explains Colman Roche, vice president of AutoStore solutions for Swisslog Americas. "Items are moving in and out faster than in a typical system."
That's because HelloFresh orders need to be filled quickly and under precise conditions to meet food freshness and safety standards. And there are a lot of moving parts: HelloFresh subscribers can customize their meal plans from more than 100 weekly menu and market item offerings and choose the day of the week they want orders delivered. The company sources meal ingredients to minimize the time between receipt and shipping, and carefully packages components to ensure freshness and simplify preparation.
The meal kits themselves may include fresh produce, starches, seasonings, recipe cards, and nutritional information along with proteins—such as fish, chicken, and beef. Subscribers can also supplement their meals with snacks, desserts, and other items from the HelloFresh Market. As a result, a typical HelloFresh outbound container includes 16 items, which must be packed in a prescribed sequence. The entire process is conducted in a chilled environment that is maintained at 33 degrees Fahrenheit.
"That insight [into the nature of HelloFresh's operations] allowed us to take a different look at the system design and come up with something that flowed more easily [in a production environment]," Roche explains.
PUTTING THE SYSTEM TO WORK
HelloFresh uses varying levels of automation at its facilities around the world, but the AutoStore system in Irving is by far the most advanced, according to DeGroot, who describes the project as a unique application of the technology. Unlike typical installations—in which orders are initiated within the AutoStore and work their way through the fulfillment process via picking stations, pack out, and eventually to shipping—orders begin outside the AutoStore, in a manual pick zone. Orders are then inducted into the AutoStore via conveyor in a sequence that maximizes throughput and worker productivity.
Roche and DeGroot describe the process as follows: Orders are initiated in a manual pick zone, where box assembly begins with proteins and ice packs. Boxed orders are then transported by conveyor and inducted into the AutoStore in a controlled sequence and dynamically grouped in batches of four based on commonalities across the orders.
The partial orders are then delivered to pickers working at carousel ports, along with bins of ingredients from the AutoStore inventory that are needed to complete the order—this allows workers to pick directly into cartons at the AutoStore ports. Workers can fulfill four orders simultaneously, maximizing the number of picks they make from each bin. Inventory is replenished through the system's 11 induction ports to maintain freshness.
Downstream from the AutoStore, orders are conveyed to automated carton sealers and labeling equipment, and then sorted for shipping.
Swisslog's SynQ warehouse execution system (WES) manages the entire process—including routing, picking across all zones, replenishment, sortation, and label printing. SynQ also provides centralized inventory management and visibility based on a first-expired, first-out (FEFO) strategy that helps ensure freshness. The WES is integrated with the facility's warehouse management system (WMS) as well.
"[SynQ] is the brain behind everything, bringing it all together," says DeGroot.
Getting the system up and running was no small accomplishment, given its size and scale. The cube-based AS/RS takes up 100,000 square feet of the facility's total 377,000 square feet of space. It includes 150 robots, nearly 30,000 storage bins, 18 carousel ports for picking, 11 conveyor ports for induction, four quality assurance ports, 6,000 feet of conveyor, four carton erectors, 12 protein pick zones with pick-to-light cells, a sortation system, and carton sealing and labeling equipment.
HUMMING ALONG
HelloFresh went live with the Irving AutoStore in 2022 and has been operating at full production since 2023. Company leaders say the system is 25% more efficient and accurate while shipping 20% more recipes compared to the rest of its fulfillment network. One of the biggest benefits is flexibility: Today, HelloFresh can easily expand its menu, adding recipes and ingredients without introducing complexity into the fulfillment process—all thanks to the software integration, synchronization, and the power of the AutoStore.
"This is something we could not have done before," DeGroot says, emphasizing the scale, speed, and flexibility of the system. "And it's the reason we use the Irving facility for expansion initiatives."
As commodities go, furniture presents its share of manufacturing and distribution challenges. For one thing, it's bulky. Second, its main components—wood and cloth—are easily damaged in transit. Third, much of it is manufactured overseas, making for some very long supply chains with all the associated risks. And finally, completed pieces can sit on the showroom floor for weeks or months, tying up inventory dollars and valuable retail space.
In other words, the furniture market is ripe for disruption. And John "Jay" Rogers wants to be the catalyst. In 2022, he cofounded a company that takes a whole new approach to furniture manufacturing—one that leverages the power of 3D printing and robotics. Rogers serves as CEO of that company, Haddy, which essentially aims to transform how furniture—and all elements of the "built environment"—are designed, manufactured, distributed, and, ultimately, recycled.
Rogers graduated from Princeton University and went to work for a medical device startup in China before moving to a hedge fund company, where he became a Chartered Financial Analyst (CFA). After that, he joined the U.S. Marine Corps, serving eight years in the infantry. Following two combat tours, he earned an MBA from the Harvard Business School and became a consultant for McKinsey & Co.
During this time, he founded Local Motors, a next-generation vehicle manufacturer that launched the world's first 3D-printed car, the Strati, in 2014. In 2021, he brought the technology to the furniture industry to launch Haddy. The father of four boys, Rogers is also a director of the RBR Foundation, a philanthropic organization focused on education and health care.
Rogers spoke recently with DC Velocity Group Editorial Director David Maloney on an episode of the "Logistics Matters" podcast.
Q: Could you tell us about Haddy and how this unique company came to be?
A: Absolutely. We have believed in the future of distributed digital manufacturing for a long time. The world has gone from being heavily globalized to one where lengthy supply chains are a liability—thanks to factors like the growing risk of terrorist attacks and the threat of tariffs. At the same time, there are more capabilities to produce things locally. Haddy is an outgrowth of those general trends.
Adoption of the technologies used in 3D printing has been decidedly uneven, although we do hear about applications like tissue bioprinting and food printing as well as the printing of trays for dental aligners. At Haddy, we saw an opportunity to take advantage of large-scale structural printing to approach the furniture and furnishings industry. The technology and software that make this possible are already here.
Q: Furniture is a very mature market. Why did you see this as a market that was ripe for disruption?
A:The furniture market has actually been disrupted many times in the last 200 years. The manufacturing of furniture for U.S. consumption originally took place in England. It then moved to Boston and from there to New Amsterdam, the Midwest, and North Carolina. Eventually, it went to Taiwan, then China, and now Vietnam, Indonesia, and Thailand. And each of those moves brought some type of disruption.
Other disruptions have been based on design. You can look at things like the advent of glue-laminated wood with Herman Miller, MillerKnoll, and the Eames [furniture design and manufacturing] movement. And you can look at changes in the way manufacturing is powered—the move from manual operations to machine-driven operations powered by steam and electricity. So the furniture industry has been continuously disrupted, sometimes by labor markets and sometimes by machines and methods.
What's happening now is that we're seeing changes in the way that labor is applied in furniture manufacturing. Furniture has traditionally been put together by human hands. But today, we have an opportunity to reassign those hands to processes that take place around the edges of furniture production. The hands are now directing robotics through programming and design; they're not actually making the furniture.
And so, we see this mature market as being one that's been continuously disrupted during the last 200 years. And this disruption now has a lot to do with changing the way that labor interacts with the making of furniture.
Q: How do your 3D printers actually create the furniture?
A:All 3D printing is not the same. The 3D printers we use are so-called "hybrid" systems. When we say hybrid, what we mean is that they're not just printers—they are holders, printers, polishers, and cutters, and they also do milling and things like that. We measure things and then print things, which is the additive portion. Then we can do subtractive and polishing work—re-measuring, moving, and printing parts again. And so, these hybrid systems are the actual makers of the furniture.
Q: What types of products are you making?
A: We've started with hardline or case goods, as they're sometimes known, for both residential and commercial use—cabinets, wall bookshelves, freestanding bookshelves, tables, rigid chairs, planters, and the like. Basically, we've been concentrating on products that don't have upholstery.
It's not that upholstery isn't necessary in furniture, as it is used in many pieces. But right now, we have found that digital furniture manufacturing becomes analog again when you have to factor in the sewing process. And so, to move quickly and fully leverage the advantages of digital manufacturing, we're sticking to the hardline groups, except for a couple of pieces that we have debuted that have 3D-printed cushions, which are super cool.
Q: Of course, 3D printers create objects in layers. What types of materials are you running through your 3D printers to create this furniture?
A: We use recycled materials, primarily polymer composites—a bio-compostable polymer or a synthetic polymer. We look for either recycled or bio-compostable [materials], which we then reinforce with fibers and fillers, and that's what makes them composites. To create the bio-compostables, we marry them with bio-fibers, such as hemp or bamboo. For synthetic materials, we marry them with things like glass or carbon fibers.
Q: Does producing goods via 3D printing allow you to customize products easily?
A: Absolutely. The real problem in the furniture and furnishings industries is that when you tool up to make something with a jig, a fixture, or a mold, you tend to be less creative because you now feel you have to make and sell a lot of that item to justify the investment.
One of the great promises of 3D printing is that it doesn't have a mold and doesn't require tooling. It exists in the digital realm before it becomes physical, and so customization is part and parcel of the process.
I would also add that people aren't necessarily looking for one-off furniture. Just because we can customize doesn't mean we're telling customers that once we've delivered a product, we break the digital mold, so to speak. We still feel that people like styles and trends created by designers, but the customization really allows enterprise clients—like businesses, retailers, and architects—to think more freely.
Customization is most useful in allowing people to "iterate" quickly. Our designers can do something digitally first without having to build a tool, which frees them to be more creative. Plus, because our material is fully recyclable, if we print something for the first time and find it doesn't work, we can just recycle it. So there's really no penalty for a failed first printing—in fact, those failures bring their own rewards in the form of lessons we can apply in future digital and physical iterations.
Q: You currently produce your furniture in an automated microfactory in Florida, with plans to set up several more. Could you talk a little about what your microfactory looks like and how you distribute the finished goods?
A: Our microfactory is a 30,000-square-foot box that mainly contains the robots that make our furniture along with shipping docks. But we don't intend for our microfactories to be storage warehouses and trans-shipment facilities like the kind you'd typically see in the furniture industry—all of the trappings of a global supply chain. Instead, a microfactory is meant to be a site where you print the product, put it on a dock, and then ship it out. So a microfactory is essentially an enabler of regional manufacturing and distribution.
Q: Do you manufacture your products on a print-to-order basis as opposed to a print-to-stock model?
A: No. We may someday get to the point where we receive an order digitally, print it, and then send it out on a truck the next day. But right now, we aren't set up to do a mini-delivery to one customer out of a microfactory.
We are an enterprise company that partners with architects, designers, builders, and retailers, who then distribute our furnishings to their customers. We are not trying to go direct-to-consumer at this stage. It's not the way a microfactory is set up to distribute goods.
Q: You've mentioned your company's use of recycled materials. Could you talk a little bit about other ways you're looking to reduce waste and help support a circular economy?
A: Yes. Sustainability and a circular economy are really something that you have to plan for. In our case, our plans call for moving toward a distributed digital manufacturing model, where we establish microfactories in various regions around the world to serve customers within a 10-hour driving radius of the factory. That is a pretty large area, so we could cover the United States with just four or five microfactories.
That also means that we can credibly build our recycling network as part of our microfactory setup. As I mentioned, we use recycled polymer stock in our production, so we're keeping that material out of a landfill. And then we tell our enterprise customers that while the furniture they're buying is extremely durable, when they're ready to run a special and offer customers a credit for turning in their used furniture, we'll buy back the material. Buying back that material actually reduces our costs because it's already been composited and created and recaptured. So our microfactory network is well designed for circularity in concert with our enterprise customers.
Managing a surging tide of post-holiday returns is a daunting task for the warehouse, where staffing shortages are all too common and inventory challenges abound. The need for speed and efficiency has never been greater.
“From the warehouse perspective, [getting] inventory back to saleable condition as quickly as possible is what really matters,” says Wes Coleman, industry principal for warehousing at mobile computing and workforce automation firm Zebra Technologies. “In this day and age of challenged supply chains, this is inventory that is readily available [and] that needs to get turned and ready for sale.”
Thanks to a growing focus on reverse logistics industrywide, more warehousing and supply chain professionals are devising new strategies and applying technology to manage the complexity of the returns cycle. Those strategies include giving retail workers a bigger role in reverse logistics and investing in software and hardware solutions to automate tasks and alleviate the physical burden of restocking items in the warehouse. Both tactics are helping organizations get a better handle on all those returns.
THINKING STRATEGICALLY
More than $700 billion worth of merchandise was returned in 2023, according to the National Retail Federation—a figure that includes an estimated 8.4 billion pounds of products that ended up in landfills. Those numbers illustrate the extent of the returns problem across the supply chain, and it’s only getting worse. Accelerating e-commerce sales and the consumer behaviors that accompany them necessitate a strong returns strategy, according to reverse logistics technology company Optoro.
“Many retailers are [rightfully] focused on the buying experience. Yet critical aspects of the customer experience happen beyond the buy button,” Natalie Walkley, Optoro’s vice president of marketing, explained in a blog post about today’s challenging returns environment. “While returns avoidance seems appealing, returns will always be inevitable to some degree, so it is wise to apply the same strategic approach to curating the best returns experience.”
For many, that means taking a step back and considering the entire returns process, which begins with the customer’s initiating a return and ends with the resale or disposition of merchandise. In the middle is the crucial step of processing the return, which happens in the warehouse. Today, many organizations are engaging front-line retail workers earlier in that cycle to help streamline the overall process.
Jim Musco, industry principal for retail at Zebra Technologies, explains that the growing complexity of returns has put the issue front and center with store operations teams—many of whom are developing presorting strategies to guide the returns journey rather than simply tossing items into a bin and sending them back to a warehouse, where pick, pack, and ship associates must figure out what to do with them. Instead, retail workers are designating whether an item can be returned to the shelves or must be sent back for further processing, as one example.
“On the retail side, they are seeing that there’s some thought that has to go into that—so that [the warehouse can get returned items in and out] quicker and more efficiently,” Musco explains. “[The industry is] becoming cognizant of the fact that the front-line retail store is part of the process—whereas before we wouldn’t have thought of it in those terms.”
APPLYING TECHNOLOGY
Technology is helping to spur that thinking. In Zebra’s “17th Annual Global Shopper Study,” released in November, nearly 90% of retail associates said they believe they can provide a better customer experience when they have mobile technology tools to help simplify communication, prioritize tasks, and check prices and inventory.
Specialized software helps as well: A returns management software (RMS) platform, for example, automates and directs the returns process. A separate 2024 returns study by post-purchase and returns management platform Narvar noted that RMS systems do several key things: enforce return windows and rules; generate a return shipping label or QR code for dropoff; route the return to the appropriate dispositioning [site] based on rules and logic; and notify customers of the status of their return. Handheld devices provide added assistance—speeding processes in the store by giving associates easy access to data and allowing them to communicate in real time via instant messages and alerts.
“These types of [technologies] are working their way into the conversation,” Musco observes. “Retailers [recognize] that reverse logistics is a big part of the equation and [that] being efficient [when processing returns] matters.”
Indeed, most retailers responding to the Zebra study said they agree that technology allows workers to do their jobs better, with 75% saying they plan to increase their technology investments in 2025.
MOBILIZING ROBOTICS
The technology trend holds true in the warehouse as well.
Some warehouses are beginning to incorporate mobile robots to help with reverse logistics, in addition to the handheld and wearable devices they already use to manage the flow of items in, out, and around their facilities. Optoro’s partnership with autonomous mobile robot (AMR) maker Locus Robotics is a good example. The companies announced a deal in 2023 to provide a scalable software and robotics automation solution for high-volume retail e-commerce returns processing. The partnership integrates Optoro’s returns technology platform with the Locus AMRs, feeding the returns information to the robots, which then navigate through the warehouse to deliver returned items to human workers for putaway in a kind of reverse picking process. The system uses data science and automated real-time decision-making to determine the best path for each returned item.
The process alleviates the physical burden associated with returns as well—human workers spend less time walking up and down warehouse aisles and more time restocking, speeding the overall returns process. Locus Robotics describes the AMR-assisted returns process as follows:
Returned goods are received by the facility.
The facility reviews and sorts returned goods.
Returned goods to be put away are sent to a holding location until the task is activated in the system.
When the task is “ready,” the returned item, container, or case to be put away is scanned and placed onto the robot, which then takes it to a human near that location.
For the human worker, the display screen shows the location, license plate, item UPC (Universal Product Code), and picture for the item or carton to be put away.
The worker puts the carton or the individual units away.
The worker places the empty carton on the robot for disposal.
The robot travels to the dropoff location. Empty cartons are disposed of, and the process begins all over again with the next item(s).
The scalable solution allows warehouses to add AMRs during the busy post-holiday returns season and scale back when volume eases.
Scalability is key, whether you’re adding robotics, mobile devices, handhelds, or wearables, Zebra’s Coleman notes.
“You want some sort of scalable option that allows you to ramp up for as long as you need to,” he says, noting that warehouses see an average return rate of about 15%, which is in line with the 2023 data from NRF.
A little help on the retail side makes a difference too.
“[There needs to be] more conversation and energy [put] into making sure you’re making appropriate decisions at the point of return or in the store as opposed to just sending [an item] to the warehouse,” Musco adds. “We’re trying to be good partners to our friends in the warehousing space.”
PITTSBURGH, PA / CINCINNATI, OH –January 23, 2025 – Freespace Robotics, an innovator in autonomous robotics for material handling and logistics, today announced a strategic partnership and investment from the Automation Solutions business of Matthews International Corporation (NASDAQ:MATW), a leading provider of warehouse automation software, controls, and order fulfillment systems. This partnership combines Freespace Robotics’ groundbreaking high-density, dynamic storage cube with Matthews’ industry-leading software, unlocking transformative capabilities for warehouse and last-mile operations.
Freespace Robotics’ pioneering solution leverages advanced robotics to deliver high-turnover, small-footprint, high-rise AS/RS (Automate Storage and Retrieval System) technologies. These innovations incorporate traditionally external conveyor functions – such as sortation, sequencing, each-picking, order buffering and pre-staging – into a seamless end to end operation. Combined with Matthews robust software suite and dashboard, which unify product information and inbound and outbound material flows, the Freespace solution offers unmatched versatility across diverse industries and specialized workflows. Together the companies are poised to strengthen core processes and improve KPIs for efficiency, scalability, throughput and cost-effectiveness.
“This partnership bridges the gap between cutting-edge hardware and proven software,” said Gary Cash, Senior Vice President and General Manager of Matthews Automation. “Freespace Robotics brings next-generation automated hardware advancements that complement and extend our software’s capabilities to achieve unprecedented order management and task efficiency in warehouse operations. Over the last decade, customers have experimented with automating some or all their logistics and warehouse operations, to varying degrees. Today they are more discerning in their choices and investments. We see Freespace as delivering on the promise of greater versatility and performance with a smart systems approach and the higher ROI customers expect.”
The collaboration will emphasize software interoperability and multi-system integration driven by customer need. Together, the companies aim to address the unique challenges of fragmented cross-docking operations and meet the specialized needs and workflows in sectors like 3PL, e-commerce, food, beverage, manufacturing and industrials.
“Software innovation has outpaced hardware advancements in warehouse automation,” said Robert J. Szczerba, CEO of Freespace Robotics. “Our system closes that gap by delivering unparalleled performance, flexibility and scalability in a more sustainable solution. Matthews’ expertise and proven software enhance our system, enabling us to supercharge automation for the most demanding industries while offering a more attractive total cost of ownership.”
Key benefits of Matthews Automation’s turn-key software suite paired with Freespace Robotics’ AS/RS:
End-to-End Task Management: Seamlessly integrates with industry-standard WMS platforms like Manhattan, Oracle, and SAP to handle inventory management, order picking, and replenishment.
Scalable, Modular Design: Adapts to diverse facility configurations, uneven floors, and unconventional layouts with high-rise options up to 70 feet. The modular rack design enables vertical and horizontal expansion to meet future business requirements, unknown demands and changing product characteristics without demolition, retrofitting, or new builds.
Cost-Effective Maintenance: Robotics and components are positioned outside the cube for easy servicing, while domestically sourced parts ensure reliability, availability and resilience.
Versatile Inventory Handling: Supports standard and large trays, full cases, and individual each-picks, accommodating loads up to 100 pounds.
Sustainable Operation: Powered by rail systems, eliminating costly and fire-prone battery dependencies while reducing environmental costs, fire risk and insurance outlays.
Shipping Buffers: Optimizes operations with seamless sorting, buffering and sequencing in a single process inside the cube for timed delivery to docks
Cross-Docking Operations: Reduces labor, equipment needs, and operational footprints while improving safety on the floor.
Last-Mile Delivery Potential: A mobile version brings goods directly to drivers in dolly order, eliminating the need to enter the trucks. This offers breakthrough performance for multi-stop delivery operations, minimizing shift time, labor and fuel costs and mid-route rescheduling.
Matthews Automation Solutions joins United States Steel Corporation (NYSE:X)(“U.S. Steel”) and 3PL leader NFI Industries as corporate investors in Freespace Robotics. Matthews’ decades of expertise and 24×7 system support, paired with Freespace Robotics’ groundbreaking technology, positions this partnership to transform logistics and warehousing operations while setting new standards of efficiency, adaptability, performance and reliability.
Freespace Robotics will also showcase its solutions at the prestigious Startup Pavilion, in booth E11200SP, during ProMat 2025, North America’s largest conference and expo for material handling and logistics professionals. The event is being held March 17-20 at McCormick Place in Chicago.
ProMat attendees can find Matthews Automation’s latest advances in Warehouse Execution System software, picking systems, and print-and-apply labeling automation at booth S4355, found in Hall A of McCormick’s South Building. Matthews’ exhibit will feature these integrated technologies – including the Freespace solution – working together to provide scalable, end-to-end applications for dynamic supply chains.
About Freespace Robotics
Freespace Robotics, a spin-out company of venture studio Carnegie Foundry, is a leading innovator in autonomous robotics solutions for the logistics and warehousing industry. By combining advanced robotics, modular design, and unmatched versatility, Freespace Robotics delivers breakthrough advances that redefine how businesses approach storage, retrieval, order fulfillment and the movement of goods. Based in Pittsburgh, PA, Freespace Robotics thrives at the intersection of industrial innovation and artificial intelligence. To learn more about Freespace Robotics, please visit www.freespacerobotics.com.
That is important because the increased use of robots has the potential to significantly reduce the impact of labor shortages in manufacturing, IFR said. That will happen when robots automate dirty, dull, dangerous or delicate tasks – such as visual quality inspection, hazardous painting, or heavy lifting—thus freeing up human workers to focus on more interesting and higher-value tasks.
To reach those goals, robots will grow through five trends in the new year, the report said:
1 – Artificial Intelligence. By leveraging diverse AI technologies, such as physical, analytical, and generative, robotics can perform a wide range of tasks more efficiently. Analytical AI enables robots to process and analyze the large amounts of data collected by their sensors. This helps to manage variability and unpredictability in the external environment, in “high mix/low-volume” production, and in public environments. Physical AI, which is created through the development of dedicated hardware and software that simulate real-world environments, allows robots to train themselves in virtual environments and operate by experience, rather than programming. And Generative AI projects aim to create a “ChatGPT moment” for Physical AI, allowing this AI-driven robotics simulation technology to advance in traditional industrial environments as well as in service robotics applications.
2 – Humanoids.
Robots in the shape of human bodies have received a lot of media attention, due to their vision where robots will become general-purpose tools that can load a dishwasher on their own and work on an assembly line elsewhere. Start-ups today are working on these humanoid general-purpose robots, with an eye toward new applications in logistics and warehousing. However, it remains to be seen whether humanoid robots can represent an economically viable and scalable business case for industrial applications, especially when compared to existing solutions. So for the time being, industrial manufacturers are still focused on humanoids performing single-purpose tasks only, with a focus on the automotive industry.
3 – Sustainability – Energy Efficiency.
Compliance with the UN's environmental sustainability goals and corresponding regulations around the world is becoming an important requirement for inclusion on supplier whitelists, and robots play a key role in helping manufacturers achieve these goals. In general, their ability to perform tasks with high precision reduces material waste and improves the output-input ratio of a manufacturing process. These automated systems ensure consistent quality, which is essential for products designed to have long lifespans and minimal maintenance. In the production of green energy technologies such as solar panels, batteries for electric cars or recycling equipment, robots are critical to cost-effective production. At the same time, robot technology is being improved to make the robots themselves more energy-efficient. For example, the lightweight construction of moving robot components reduces their energy consumption. Different levels of sleep mode put the hardware in an energy saving parking position. Advances in gripper technology use bionics to achieve high grip strength with almost no energy consumption.
4 – New Fields of Business.
The general manufacturing industry still has a lot of potential for robotic automation. But most manufacturing companies are small and medium-sized enterprises (SMEs), which means the adoption of industrial robots by SMEs is still hampered by high initial investment and total cost of ownership. To address that hurdle, Robot-as-a-Service (RaaS) business models allow enterprises to benefit from robotic automation with no fixed capital involved. Another option is using low-cost robotics to provide a “good enough” product for applications that have low requirements in terms of precision, payload, and service life. Powered by the those approaches, new customer segments beyond manufacturing include construction, laboratory automation, and warehousing.
5 – Addressing Labor Shortage.
The global manufacturing sector continues to suffer from labor shortages, according to the International Labour Organisation (ILO). One of the main drivers is demographic change, which is already burdening labor markets in leading economies such as the United States, Japan, China, the Republic of Korea, or Germany. Although the impact varies from country to country, the cumulative effect on the supply chain is a concern almost everywhere.
Picture a busy DC, with manually operated forklifts, people, and pallets in constant motion. At the same time, the stationary equipment they interact with, such as conveyors and palletizers, is industriously whirring away. Together, they are performing something akin to a carefully choreographed ballet.
Now add driverless forklifts to the mix. Shuttling along without a human operator on board, they may look like they’re operating independently, but they’re not. They’re actually in constant contact with other equipment and software, making sure they perform their part in the dance at the right moment. Without that ability to communicate, the forklifts—and other warehouse operations—could come to a standstill.
Who, exactly, are driverless forklifts “talking” to, what information are they sharing, and how does that exchange happen? We asked automation experts to explain. They also shared tips on ensuring successful communication between automated lift trucks and other equipment and software.
TWO-WAY COMMUNICATION
Lift trucks that do their jobs without a human operator on board cannot “speak” directly to each other. “As it stands now, there is no peer-to-peer communication or interaction on a forklift-to-forklift basis,” notes David Griffin, chief sales officer for Seegrid, a developer of autonomous lift trucks and AMRs (autonomous mobile robots). There is, however, interaction between forklifts via a centralized fleet manager system (also referred to as a traffic management system or an automation server). This “overarching conductor of the automated system” assigns tasks to each forklift, controls the route the trucks will follow, and manages traffic flow, says Nick McClurg, a sales engineer at forklift maker Hyster Co.
The forklifts communicate with many kinds of material handling equipment, such as robotic palletizers and depalletizers, stretch wrappers, conveyors, automated storage and retrieval systems (AS/RS), and dock equipment. That communication must be bidirectional, says Michael Marcum, senior director of autonomous vehicles at systems integrator Bastian Solutions, a Toyota Automated Solutions company that also makes robotic forklifts. Much of the exchange consists of messages that indicate status—whether or not the two pieces of equipment involved are ready to conduct a transaction. For example, if a forklift will be delivering a pallet to a stretch wrapper, then the wrapper has to tell that forklift, via the fleet manager system, that the load position is empty and the forklift is allowed to set a payload there, Marcum explains. After a pallet has been wrapped, the stretch wrapper will call for a pickup via the fleet manager. Once the forklift picks up the wrapped pallet, it must confirm to the stretch wrapper that it has departed; without that signal, the wrapper cannot receive its next load.
If a truck is not ready for an assigned task, it signals that status to the fleet manager, and the task will be reassigned to another nearby vehicle, says Jayce Nelson, sales manager, North America, for Kion Group’s Linde Automated Solutions, a specialist in automated forklifts and software. When the assigned forklift is ready to approach, say, the end of a conveyor to pick up a load, it uses its vision systems, such as 3D cameras, to align itself with the equipment.
With their control software, robotic forklifts also have the ability to communicate with other warehouse equipment, like fire-detection systems and automated rollup doors. “If a device is capable of sending or receiving electrical signals, then the vehicle can interact with that device via the automation host software,” McClurg says. Even a piece of mechanical equipment could be outfitted with sensors that help it interact with automated forklifts, according to Brian Markison, director of sales for Rocrich AGV Solutions, a joint venture of Mitsubishi Logisnext’s Rocla and Jungheinrich units that specializes in automated guided vehicles.
The capability to communicate with different types of devices enhances warehouse safety, Griffin says, because it enables automated forklifts to talk to safety equipment like pedestrian warning lights and intersection gates. And since the robots constantly transmit their location, the traffic control system can identify developing problems and prevent them. For example, the system will stop an autonomous forklift from entering an intersection that’s occupied by another lift truck. Once the other truck has moved on, the system will give the approaching forklift the “all clear,” he says.
Hardware isn’t the only thing driverless forklifts can talk to; they also are in continual dialogue with various types of software. “Most commercial warehouse software programs today have the capability for two-way communication, and most can be integrated with automated lift truck fleet management software,” observes John Wilkins, a sales engineer for Yale Lift Truck Technologies. The most common are warehouse management systems (WMS) and warehouse control systems (WCS); others include enterprise resource planning (ERP) software, fleet management and telematics systems, and transportation management (TMS) and order management (OMS) systems.
As for how that might work, Rocrich’s Markison gives the example of a WMS sending an order to move a pallet from one location to another. The order typically will include start and completion time, and some indication of the move’s priority. “That order can then be taken into the fleet manager, which will appropriately queue up the tasks that need to be done,” he explains. The forklift must also report completed missions back to the WMS.
HOW TO TALK TO A FORKLIFT
Communication between robotic forklifts and warehouse equipment and software happens in a number of ways. Which method is deployed depends on the equipment and software involved as well as the tasks to be carried out. Each installation is unique in some way, but there are some commonly used approaches.
Some communication protocols are more widely used than others. Examples of those in widespread use include modbus, a serial communication protocol that governs an initiating and a responding device, and CANbus (Controller Area Network), a real-time communication protocol that transmits data to networked industrial controls.
A driverless forklift’s interface with other equipment could be something electromechanical, such as a photo-eye sensor, says Jeff Kuss, product manager–automated solutions at forklift maker and intralogistics specialist The Raymond Corp. A sensor at the end of a conveyor, for instance, could detect the presence of a pallet. That triggers the sensor to create an electrical signal that it sends to a programmable logic controller (PLC). The PLC receives the electrical signal as a digital input and then transmits a message, via ethernet, to the server that controls the automated vehicles. Finally, the server sends the instructions over Wi-Fi to the closest available forklift to “pick up the pallet and take it to Location X.” (Some facilities use Bluetooth or cellular transmission instead of Wi-Fi.) Data that identify loads and trigger a task can also be acquired through IoT (internet of things) platforms, RFID (radio-frequency identification) systems, and barcode scanning.
Another option, Bastian’s Marcum says, is to use infrared-based optic couplers that share bits (binary digits, the smallest units of digital information) as inputs and outputs. When the forklift gets within a certain distance of another piece of equipment, “the two devices can talk to each other, similar to the way a TV remote works,” he explains.
Usually, though, software is a critical intermediary between driverless forklifts and other equipment. It can be complicated. In the case of a WMS, McClurg says, his company’s approach is to send a text or JavaScript Object Notation (JSON) file to the WMS; in exchange, the WMS sends a file to a folder on the localized network that can be accessed by the automation host software. The fleet manager reads the file and executes it. Once the task has been completed, a message is sent to another folder. The WMS opens it, reads it, and, based on its contents, either closes out the order or sends additional instructions.
To ensure that interfacing software programs understand the messages they receive from each other, it’s often necessary to create an application programming interface (API). An API is a comparatively simple type of middleware—or software layer—that acts as a translator, facilitating communication by reformatting messages so they will be intelligible to the receiver. In essence, they are “setting ground rules in terms of what information is passing back and forth and what it means,” Markison explains.
In some cases, more complex middleware may be needed. According to Brice Bucher, senior manager of products at software developer and systems integrator Flexware Innovation, APIs have limitations. In a presentation at the Autonomous Mobile Robotics & Logistics Conference 2024, Bucher noted that APIs don’t address data transformation, protocol conversion, or business logic integration. When each system has different data formats or requires specific protocols, middleware bridges those gaps, he said. Middleware also ensures that data moves between systems without delay, he said. For example, if an AGV completes a task, middleware can instantly trigger updates across systems, so that WMS, ERP, and other systems are aligned in real time.
CAN WE TALK?
Raymond’s Kuss notes that each communication integration will be unique in some way. That’s partly because automated forklift vendors and suppliers of fleet manager systems have proprietary interfaces. On top of that, software with some degree of customization, such as a WMS, may require modifications to the fleet manager system, he explains. What’s more, adds his colleague John Rosenberger, director, iWarehouse Gateway & Global Telematics, “even if we know the format for efficient data transfer, the content of the messages may differ depending on the forklift manufacturer, or it can be different by functionality.”
Mixed fleets with forklifts from different manufacturers present a particular challenge. Seegrid’s Griffin notes that it’s common for facilities to use robots from multiple vendors. Generally, he says, each automated solution has its own proprietary fleet manager software that understands where all units under its purview are and controls their movements. When robots of different brands cannot be confined to separate areas, it’s important that their fleet managers have the ability to communicate, so they can do things like open and close intersections where different types of robots cross paths.
While it is possible for dissimilar fleet managers to talk to each other, that’s easier said than done. “Those systems inherently are not interoperable,” Nelson says. “The need to share information like coordinates, current status, past assignments, and prioritization makes it difficult to assign travel paths.” In addition, if the forklifts are unable to communicate location information and what they are doing, that can lead to deadlock, where the vehicles simply stop—what Yale’s Wilkinson calls “the classic situation: a staring contest between two autonomous vehicles from different OEMs, neither one capable of blinking or losing.”
A solution for some facilities is third-party fleet manager software that’s designed to work in multiple brands of autonomous forklifts; examples include those offered by independent developers such as Kollmorgen, BlueBotics, Navitech Systems, and Flexware Innovation. In fact, some forklift OEMs partner with these and other providers instead of developing their own fleet managers. This opens the way for a fleet to potentially buy different robots utilizing the same control and navigation system, which will reduce complexity to some degree, Marcum says.
Communications with driverless forklifts may become simpler in the near future. VDA 5050, an open-source protocol for communication between AGVs and fleet manager systems, is currently in development. Coordinated by two German industry organizations, one for auto manufacturers and the other for material handling and intralogistics, this universal protocol promises to allow “any mobile robot, regardless of brand, [to] be seamlessly integrated into existing operations,” wrote Alfredo Pastor Tella, who runs the Europe-based AGV Network website, in a LinkedIn post. Pastor Tella wrote that Kollmorgen will introduce VDA 5050 into its robot control software in 2025, but other industry observers have noted that because the protocol’s roots are in European manufacturing and there are still technical issues to be worked out, it may be a few years before it takes hold in the forklift world. When it does, conversations with autonomous forklift fleets will likely become much less complicated to hold.
Tips for success
Want to be sure your driverless forklifts will always “get the message”? Here are some experts’ recommendations for facilitating communication with them:
Involve your IT experts early! They’ll need to identify what relevant data is currently available and where it resides. Make sure they’re comfortable that any APIs and other software meet your company’s security requirements. For cloud-based systems, verify that the vendor and systems integrator will have remote access if they need to service any of the systems or software. (Brian Markison, Rocrich AGV Solutions)
If you’re buying from different manufacturers, find out which supplier has navigation technology on the brands you’re considering and try to stay with a single system if possible. If you have a single platform, you can make a change just once and the entire fleet will receive that modification. If you have two fleet managers, segregate them as much as possible. Wherever they are separate, you’ll only have to change that one, but in shared areas, you’ll have to change both. (Michael Marcum, Bastian Solutions)
When it comes to facilitating communication, software is not always the best answer. Sometimes something simpler, like PLCs that notify equipment through very basic logic, works just fine. And it’s better to start small and integrate each function as you go, rather than try to integrate everything at once. You can tie two systems together and demonstrate the benefits from that, then use the savings to justify and help fund the next piece. (John Rosenberger, The Raymond Corp.)
Conduct testing in real-world scenarios, and make sure legacy software and communication technologies are compatible with the automation. These systems work in a dynamic environment, and a lot changes over time. Calibration tests can make sure everything still aligns correctly. And remember to inform your vendor of changes in things like throughput rates, layout, pallet sizes and configurations, products, and so on. (Jayce Nelson, Linde Automated Solutions)
In most facilities, commands and data are communicated via Wi-Fi, so connectivity and reliability are a top concern. A pre-installation survey to measure Wi-Fi signal availability and strength throughout the facility is an absolute must. Based on those findings, you may need to enhance signal strength and expand capacity and coverage. In some very large facilities, a private wireless network that uses cellular signals may be the best solution. (Deryk Powell, CEO, Velociti Inc., a provider of technology deployment, support, and integration services)