As autonomous mobile robots (AMRs) assume increasing importance in the supply chain, more and more vendors are emerging to build these robots to complete specific tasks. Each of these companies’ creations is unique—based on their own distinctive language, algorithms and software. That makes coordination among them almost impossible.
But, imagine if robots could talk to other robots—no matter where or how they were designed and built?
Many in the field of AMRs believe that interoperability among robots will be crucial to more widespread adoption of the technology in factories, warehouses and distribution centers. Interoperability would allow robot fleets from different vendors to coordinate activities, share data and adopt the same operating conventions.
Toward that end, MassRobotics, a nonprofit focused on robotics solutions, announced the release of open-source interoperability standards for AMRs in May. The standards, which were created by an industry working group, will be tested at a FedEx warehouse in Tennessee this year with robots from multiple vendors.
Jason Walker, CEO and co-founder of MHI member Waypoint Robotics, one of the participants in the creation of the standards and the initial test with FedEx, said interoperability’s importance is only going to grow, and the standards represent an essential step toward making interoperability achievable.
“I do believe that everyone in the community recognizes how important interoperability is, and how detrimental it could be if we didn’t address it early,” Walker said. “And the fact that so many key players stepped up and made it a priority to establish a standard in cooperation with MassRobotics, I think it’s a really big deal.”
‘A closed loop’
As it currently stands, AMRs are running on “a kind of closed loop” that prevents coordination among multiple vendors in a facility, said Melonee Wise, CEO of MHI member Fetch Robotics.
“It gets a sensor signal, it does what it’s supposed to do, and it continues doing it … but all of that changes when you want to have System A talk to System B,” Wise said. “Now, the input is coming from some other system, and that’s where you see a lot of challenges today.”
Jeff Christensen, vice president of product for MHI member Seegrid, said robotics are complex enough that no single vendor will ever have a robot for every task.
“It’s untenable,” he said. “And since customers are buying a solution to a business problem, and not a set of robots just because robots are cool, that means that they need to be able to apply the technology to the business problem in the way that makes the most sense for the business. That will mean different vendors, and that will mean interoperability.”
“It is inevitable that everybody who has enough diversity in the needs of their AMR, they’re going to hit this wall,” added Walker. “You have robots with operational parameters that are relatively narrow. and that means that you’ve got all kinds of different tasks that need to be solved.”
The wide scope of needs means there’s a variety of options when it comes to solutions. “Everybody’s trying to solve a lot of different kinds of tasks in different ways. And pretty much all of them are useful in one way or another,” said Walker. ‘So, because it is necessary to have heterogeneous fleets, it’s necessary to have interoperability between those fleets. If you’re the customer … and you need a fork robot, a deck load robot, a scanning robot, and a pallet-moving robot, then being able to have one system that can work with all of those other systems, and also have those other systems interact and coexist in the same facility together in a safe and productive way—everybody in the community really thinks that’s going to be the next choke point.”
A fully connected future
Interoperability offers an array of potential benefits, said Zachary Dydek, CTO of MHI member Vecna Robotics.
“To me, it looks like a warehouse of the future that has robots from many different vendors, as well as people, all doing the things that they are good at, respectively, in a coordinated way,” Dydek said. “And they’re not just coexisting, but actually coordinating with each other to drive the efficiencies that we need in the supply chain right now to deal with increasing consumer demand.”
Chris Caccioppo, CTO and cofounder of MHI member 6 River Systems, said interoperability is the key to operating an efficient and successful automated warehouse system.
“Warehouse systems and the AMRs’ software need to be able to share data between them, but those systems are not all the same and may have different data formats,” Caccioppo said. “Interoperability standards solve this challenge by allowing applications to seamlessly integrate, sharing data, such as inventory data, shipping information and purchase orders. It should be flexible enough to adapt to changing business models, satisfy operational requirements and keep pace with industry best practices and regulatory compliance requirements.”
In addition, Caccioppo said interoperability will require less human assistance or intervention when system changes occur.
Interoperability standards represent a prerequisite to bigger things, such as “fully connected solutions,” Christensen said.
“If you think of the larger solution of connected devices, connected robots, connecting your islands of automation together in order to achieve a larger goal, [standards are]kind of a prerequisite before you get to that loftier goal,” Christensen said. “But that’s really where it becomes really, really interesting is that if robots today can automate moves of material, a system of coordinated robots that understand the operation can eventually automate the continuous improvement of the moves of the operation.”
Walker believes there has been a reticence for some companies to “lean into” AMR because they are waiting to see which AMR solution emerges as “the winner.” However, he said, the industry is one that will have many winners rather than just one. Interoperability, he believes, will help companies overcome their wariness to committing to one technology or company.
In that vein, Wise believes some of the interest in interoperability stems from interest in “exchangeability” or “commoditization.” Rather than centering on the question of how to get two products working together, those concepts are concerned with how to seamlessly replace one vendor’s product with another from a different vendor. Wise said she sometimes hears from customers interested more in commoditization than interoperability.
“The conversation often is couched in this notion of robots working together, when sometimes the question is really not about whether your robots work together, but, ‘Can I replace Robot A for Robot B?’” Wise said.
Wise said the prospect of commoditization appears further away than interoperability.
“It’s a really hard question that a lot of people want answered. but I don’t think that there’s a good answer for it,” Wise said. “In the same way that it took almost 30 years to commoditize PCs, it will take even longer for us to commoditize robots, because so much of the thing that you’d be trying to commoditize is the probabilistic software that we’re developing—it’s not the hardware.”
Dydek said the standards are just one step, but an important step, in the march toward interoperability and coordination. For standards to ultimately work, Dydek said, robotics vendors will have to embrace them. That could be driven by customers, he said.
“There has to be a kind of a pull from the customer, the end user of the robotic system,” Dydek said. “They’ll need to say, ‘Hey, I need these robots to interoperate. I’ve got these different vendors in the house, and they’re not working together the way I need them to. I need to be able to do these coordination-type activities better than they’re doing it right now.’ And we’re already seeing some of that … We’re seeing that sort of shift in the minds of our customers. And that’s going to really accelerate things.”
Wise said a central challenge to interoperability is the question of “How do you explain the world to the robots, so that the description is the same from one provider to the other?” Getting robots to interoperate will depend on them having “the same semantic view of the world,” she said, so that instructions for tasks are clear to all of them.
“Task specification and what you’re able to do in the world is very dependent on the language that you have to define the world,” Wise said. “If you can’t all agree what it means to go to someplace, then you will have a very hard time telling all the robots to go there.”
“If the robotic solution that a fulfillment center is planning on implementing is not compatible with their current WMS, it can create roadblocks such as an interruption in workflow and/or requiring management and staff to learn a new system, thus increasing training time,” Caccioppo said.
In an optimistic time-horizon, Wise believes “we’re five to 10 years from creating interoperability between AMRs for the point of doing interactions between different platforms.”
Christensen agreed that interoperability is “a future play.”
Fully autonomous robotics “is being applied to specific kinds of applications, and customers are getting great returns for that,” Christensen said. “That’s great, and there will be a natural evolution toward things that are more integrated. But you can go a long way with single-vendor solutions for a while. So, I think that [interoperability is]not a critical blocker right now.”
For customers who have leaned into AMRs already, Walker said, they know they will need a heterogeneous fleet to expand automation and that interoperability needs to be solved to make it possible. The MassRobotics standards are a sign of what’s to come.
“The fact that that standard exists, it gives everyone something to aim at,” Walker said. It does not mean immediate interoperability, but “it gets everybody pushing the rock in the same direction,” he said.
Christensen said building consensus and commitment around standards will take time. He noted that the large number of competing vendors in robotics—a number that is only growing—means that it will be an ongoing process.
“The benefits to a customer will be real, but they’re still a few years out into the future,” Christensen said. “There’s lots of disparate innovation that needs to coalesce around some baseline of interoperability, and then that baseline evolves over time until there’s kind of a critical mass around it. That’s how these things evolve. And it’s happened lots of times before, and I anticipate it’ll happen for robots as well.”