Demand for piece-picking robots gains steam

Piece-picking robots are finding their way into more warehouses and distribution centers (DCs) these days thanks to steady improvements in the technology that are making it more attractive to a wider range of end-users. Advances in gripping technology and arm speed are making it easier to handle a broader array of items, for example, allowing companies to maximize their investment and reduce their reliance on human labor for mundane picking tasks. 

The trend is part of a growing demand for industrial robots in general—a trend that is only expected to gather steam as fulfillment operations deal with rising e-commerce volumes and the labor challenges that accompany such growth. The Association for Advancing Automation (A3) tracked a 28% increase in North American robot sales last year compared to 2020; it was the strongest sales year on record, with $2 billion worth of robot units sold. Separate industry statistics valued the piece-picking robot market at more than $148 million in 2020, a figure that’s expected to surpass $3 billion by 2026, a nearly 63% compound annual growth rate. 

The growth is spurring interest from the investment community as well. Earlier this year, robotic picking solutions developer RightHand Robotics landed $66 million in venture capital funding, which the company said it will use to accelerate the development of its RightPick piece-picking solution. There’s been a steady stream of similar announcements regarding robotics-industry investments over the past year, which underscores the bullish market outlook for the technology.

“We foresee strong growth in this area,” explains A3’s President Jeff Burnstein, emphasizing e-commerce growth and a related demand to build more warehouses and DCs—all of which will need automation. “That means this segment will grow, and there will be innovation. There is a lot of venture capital [flowing] into this space as well, which shows you this will be a very aggressive market moving forward.”

Some of the most common logistics applications for piece-picking robots today are sortation and pick-and-place functions in general merchandise, apparel, and small-package operations, but experts say the opportunities are growing just as fast as the demand.

GETTING BETTER OVER TIME

There have been “massive improvements” in piece-picking technology over the past few years, according to Jake Heldenberg, senior manager for warehouse solutions sales consulting at material handling systems integrator Vanderlande. Piece-picking robots use a gripper attached to a mechanical arm for item picking and are most commonly used to pick single items from a source bin and deposit them into an outbound container. Heldenberg says grippers are able to manipulate and grasp far more items than they could just a few years ago and that arm speed has improved as well, allowing companies across many industries to meet productivity improvement goals. 

“Four years ago, [robotic piece-picking solutions] could handle 60% to 70% of SKUs [stock-keeping units]. Now, they can handle 90%-plus for picking in general merchandise and fashion,” Heldenberg says. “The most successful products have been cosmetics—because they come in small boxes that are lightweight and easy to grip.”

The biggest challenge has been the robot’s vision system, but that technology is improving as well. Piece-picking robots incorporate three-dimensional (3D) cameras and software to “see” what they are doing. Robot developers are working to improve vision systems so that robots can more easily identify items of different shapes, sizes, and weights. Artificial intelligence (AI) and machine learning (ML) can help with this process by allowing the robot to “learn” and improve on its own. 

“Maybe the robot is being told to pick object ‘X’ out of a bin of various objects. To do that, it has to identify object ‘X,’” Burnstein explains, adding that understanding the item’s dimensions and weight are a crucial part of that process. “Robots can learn how to do this more effectively over time through AI and machine learning.”

Vince Martinelli, head of product and marketing at RightHand Robotics, explains that AI gives the robot the cognitive skills and “understanding” of its space that allow it to function more accurately and consistently in a complex environment. He says today’s piece-picking robots are more adaptable, reliable, and faster than ever before—thanks to advances in AI, but also because developers are gaining real-world experience as their products become more widely used. Real-world applications can reveal obstacles, errors, and scenarios in which a robot may not complete a task perfectly, for instance. Developers can then use that experience to further refine the technology’s capabilities. 

“Some things are hard to do in a lab,” Martinelli says. “We can’t pre-imagine every scenario. If something isn’t perfect, how do you resolve that in the field? Learning how to collect, collate, and process the data coming from the machines [helps us] relentlessly drive overall reliability and performance.”

REALIZING SUCCESS IN THE FIELD

Piece-picking technology is becoming an increasingly important part of the automation mix for parcel carrier and logistics services provider FedEx, according to Aaron Prather, senior adviser for the company’s technology research and planning team. The company uses piece-picking robots for sortation and pick-and-place operations at facilities around the world, and plans to use them for even more applications at both new and existing facilities—thanks in large part to technology advances that have spurred creativity throughout the organization. Over the past five years, the company’s FedEx Ground business has added more than 60 automated stations to its network, with nearly 150 fully automated facilities in the ground network, which affect more than 97% of package volume.

 “Right now, the technology is good at picking up small packages and placing them on a [conveyor] belt to go into the system,” Prather says, explaining that robots are replacing humans at drop-off facilities and larger FedEx warehouses where parcels and small packages are sorted, scanned, and sent on to their next destination. “But each day, we’re looking at doing something else [with the technology]. There is a lot of growth in this area. There’s nowhere to go but up.”

Prather says next steps for robotic piece picking include unloading trucks—another pick-and-place application where the company could free up human labor for other activities. The idea is to apply the technology at warehouses and DCs in the network as well as at airfreight facilities, where parcels and packages must be unloaded from large containers. 

“That is the next big type of application a lot of us are going after,” Prather says, adding that the robots would pick parcels and then place them on a conveyor belt or possibly an autonomous mobile robot (AMR). “If you’re in logistics, that’s something you want to solve—it’s a labor-intense activity, so it’s something that everyone wants to automate.

“There are so many interesting use cases coming up now, because the technology is getting so much better.”

He says the technology is also spurring a workforce evolution that is creating higher-level career opportunities throughout the organization. FedEx’s new “robot team leader” positions are a case in point. These are hourly jobs in which an employee oversees a group of robots and is trained to manage, monitor, troubleshoot, and address any maintenance issues that may arise. 

“Some are watching a bunch of [robotic] arms; others are watching a bunch of mobile robots. They are trained on robot support,” Prather explains, emphasizing the importance of the position as well as a growing interest in it among employees. “All of them love their jobs. They get to go home and tell their kids they work with robots. But we can’t create more robot team leaders until we put more robots in.”

And that’s in the works. Most recently, FedEx installed a sorting robot at a FedEx Express facility in China that handles small inbound and outbound packages for e-commerce customers in the southern part of the country. The project followed similar sortation solutions implemented here at home last year. One example is the implementation of robot developer Berkshire Grey’s Robotic Product Sortation and Identification (RPSi) system at FedEx Ground facilities in New York, Las Vegas, and Ohio. The AI-based system autonomously picks, identifies, sorts, and collects small packages that were previously sorted by hand.

That kind of innovation will only continue, he says.

“We know going forward that our greenfield sites will have automation. We’ll design sites with this technology in mind,” Prather says. “However, it is still critical for the industry to understand there are a lot of brownfield sites we are not going to give up on. Our volume continues to grow, people are shipping like crazy, so we are still going to look at ways to automate those sites as well. The challenge will be how do we take these technologies and fit them into these brownfield sites? That’s where our creativity is going to kick in.”