The December 2022 issue of IEEE Spectrum is here!

Close bar

5G Is Putting Robots’ Heads in the Cloud

Low-latency 5G networks could help solve a “critical problem” for smart robots

5 min read
Dal.Komm Coffee's Beat robot provides caffeine sustenance to attendees of MWC Barcelona.
Photo: Michael Koziol

Dal.Komm Coffee's Beat robot provides caffeine sustenance to attendees of MWC Barcelona.Dal.Komm Coffee’s Beat robot provides caffeine sustenance to attendees of MWC Barcelona.Photo: Michael Koziol

I almost missed out on the vanguard of the 5G robot revolution because I don’t drink coffee. The near miss occurred on Monday at mobile network trade association GSMA’s booth at MWC Barcelona (formerly called Mobile World Congress), where Dal.Komm Coffee was demonstrating a coffee-serving robot. Thankfully for non-coffee drinking weirdos like me, hot chocolate was also available.

After placing an order via a smartphone nearby, a robotic arm behind a glass panel juggled cups, operated coffee makers, and gently placed drinks on trays so they could be collected by waiting humans. According to a representative of Dal.Komm, the robot’s precise movement was only possible with a 5G network provided by the KT Corporation, the Korean teleco.

The 5G network supports a pair of cameras on the ceiling of the robot’s enclosure. One looks outward, keeping an eye on waiting customers, while the other peers inward, watching the robot’s environment. From that second camera, the robot has a real-time 3D picture of its own surroundings, without which it couldn’t manipulate cups and juggle orders as precisely as it did.

As 5G networks begin to roll out around the world, the lingering question that I’m exploring this year at MWC Barcelona is—what will we actually use these ultra-fast, low-latency networks for? Telecos have invested billions of dollars and the better part of a decade in bringing high-speed 5G networks to customers. Though there are many possible applications of this technology, for both consumers and companies, it’s not yet clear which ones will catch on.

For those that do, 5G is often described as an enabling technology that will spur further developments in industries and sectors that figure out how to leverage its speed and bandwidth for true commercial gain, rather than trade show gimmicks. And based on what I’ve seen so far at the show, robotics is one of the industries that could benefit most from 5G service.      

Back to the hot chocolate. The robot itself—called Beat and referred to as a “robot cafe”—is already serving drinks at 40 locations in Korea. But those robot cafes, which lack a 5G connection, are not as precise in their movements, cannot move as fast, and cannot handle as many orders as the second generation of cafes now on display at MWC Barcelona.

Plenty of this year’s demonstrations show how 5G, the next generation of wireless, could allow robots to do more tasks, and do tasks more intelligently. In many cases, including for Beat, 5G’s high data rates (offering peak download speeds of 10 Gbps versus 1 Gbps on today’s networks) will allow robots to wirelessly transfer the immense amount of data involved with video surveillance or machine learning processes.

What’s even more important, however, is the low latency 5G offers—expected to be 10 milliseconds or less versus roughly 60 milliseconds on 4G. Without this improvement, wireless robots can’t quickly act on information they gather. The difference between a robotic barista knocking over or picking up a cup of coffee relies in part on how quickly its brain can tell the bot to stop moving its arm.

SK Telecom, another Korean teleco, is exhibiting a smart factory where robots scan products on an assembly line for defects. The robots rapidly take dozens of photographs of each item and crunch the data quickly—transmitting it over 5G to the cloud for processing—to spot imperfections. In the demo, robots checked pucks for minute scratches, many of which were difficult to spot with the naked eye. Robotic arms then sorted the pucks into one of two conveyors, based on whether or not they had passed the inspection.

[shortcode ieee-pullquote quote=""I realized the robotics problem is not robotics. It's a communications problem."" float="left" expand=1]

It’s also clear that engineers are thinking plenty about the intersection of these more capable robots and their human supervisors. Elsewhere in the expansive conference halls, Intel is showing how 5G can help keep humans safe from their metal coworkers.

In the company’s demonstration, a section of floor was marked as a danger zone. Cameras monitored the area, watching for people who wandered in. When someone did enter, a nearby robot was immediately signaled to stop its work (admittedly, its work was not particularly dangerous, as it was folding paper airplanes).

A representative from Intel explained that the cameras were using their own AI systems to determine when they needed to tell the robot to shut down (with those instructions transmitted immediately over 5G).

But to really get a taste of what 5G can offer robotics—and for what’s to come—there’s Cloudminds. The company designs and builds robots with AI brains in the cloud, which connect to the body thanks to a 5G network. If that seems overly complicated—"Just put the robot’s brain in its body,” you might exclaim—Cloudmind’s CEO and founder, Bill Huang, thinks it’s the only option: An electronic brain complex enough for the robot to be capable of solving complicated problems would have to be far larger, and heavier, than any human-sized robot could hold. According to Huang, the brain’s complexity is simply too great to reproduce it in silicon. “That’s why God didn’t use electronics,” he says.

Daisy, a 5G cloud robot from Cloudminds.Photo: Michael Koziol

Instead, the company’s XR1 service robots—all of which, it’s worth pointing out, had traditionally female voices and names like Rachel, Daisy, and Nicole—may have their physical bodies in an exhibition hall, but their brains are in the cloud, distributed in servers around the world. And, as I’m sure is no surprise at this point, 5G is keeping their brains and bodies in sync.

With that 5G connection, each XR1 uses camera vision to observe its surroundings and manipulate its limbs precisely enough to serve drinks or thread needles. Not content to stop at the brain, however, every actuator in the XR1 also has its own IP address. In effect, every joint in the robot is a node in a mesh network, making the robot its own 5G network—capable of precisely controlling each joint for smooth and accurate movement.

The XR1s can learn from their own experience, but they can also learn from having a human operator take remote control of their body—again, 5G helps make this possible—to perform a task. To learn to thread a needle, for example, a human can take control, walk the robot through the motions, and then allow the robot to perfect its technique through practice.

The XR1 robots build upon work earlier seen with Pepper, a robot that has been charming humans with conversation since 2014—and officiating funerals since 2017. A representative from Cloudminds explained that Pepper’s conversational skills work over a 4G network—the data throughput and low latency can handle language processing—but it’s not enough for the kinds of skills the XR1s are capable of. Pepper has human-like hands, but cannot process data quickly enough for them to be useful and precise.

Huang is excited about the possibilities that 5G opens up for robotics. In 2012 while he was at China Mobile, he says he could see what was holding robotics back. “I realized the robotics problem is not robotics,” he says. “It’s a communications problem.” He founded Cloudminds in 2015 to pursue his idea of treating the robot’s body as a bunch of actuators and putting its mind in the cloud. Until 5G, however, the idea was still limited by the latency of 4G networks. “5G solves the critical problem,” he says.

The Conversation (0)

Why the Internet Needs the InterPlanetary File System

Peer-to-peer file sharing would make the Internet far more efficient

12 min read
An illustration of a series
Carl De Torres

When the COVID-19 pandemic erupted in early 2020, the world made an unprecedented shift to remote work. As a precaution, some Internet providers scaled back service levels temporarily, although that probably wasn’t necessary for countries in Asia, Europe, and North America, which were generally able to cope with the surge in demand caused by people teleworking (and binge-watching Netflix). That’s because most of their networks were overprovisioned, with more capacity than they usually need. But in countries without the same level of investment in network infrastructure, the picture was less rosy: Internet service providers (ISPs) in South Africa and Venezuela, for instance, reported significant strain.

But is overprovisioning the only way to ensure resilience? We don’t think so. To understand the alternative approach we’re championing, though, you first need to recall how the Internet works.

Keep Reading ↓Show less