BOSTON — The opening and closing speakers at last week’s Robotics Summit & Expo here both focused on how robotics developers can prepare for 5G service. The next-generation ultra-wideband networks will reduce latency and increase data throughput and reliability, enabling new robotics applications in the next few years, said the keynoters.
First-generation wireless technology refers to analog cellular phone service, which began in the 1980s. Joshua Ness, senior manager at Verizon 5G Labs in New York, recalled “brick” phones in his breakfast session. 2G refers to digital cellular service, based on GSM, the Global System for Mobile Communications. It allowed text messaging.
3G started in 2000 and provided the bandwidth for the first smartphones, including streaming video. From the 2010s onward, 4G has provided a download speed of 100Mbps for higher-quality streaming video, better security, and wider service. “The gig economy was built on 4G,” observed Ness. He added that Verizon’s 4G node for the U.S. Northeast is in Yonkers, N.Y.
Over the past several years, telecommunications providers, chip manufacturers, and robotics and artificial intelligence companies have begun exploring how 5G will enable the “wireless World Wide Web” and computing at the edge rather than in the cloud.
Verizon 5G mobile service is available in Chicago and Minneapolis, and fixed home broadband is starting in Houston, Indianapolis, Los Angeles, and Sacramento, Ness said. Boston is on the list of U.S. cities to get 5G soon, he added.
Verizon encourages 5G collaboration
“5G enables supporting technologies, such as computer vision and edge computing,” said Ness. “With software-defined supercomputing resources pushed out to nodes, which you’ve seen everywhere as utility boxes, you’re not limited by distance to the cloud.”
This is important for the Industrial Internet of Things (IIoT), as well as reducing the need for compute resources and thus power and battery onboard phones. With speeds of 1Gbps, 5G will also shorten processing time for applications such as self-driving cars and robots, he said.
“What that means from a hardware perspective is that you can offload compute from the device, changing the design constraints,” Ness explained. “This can also lead to AR and VR [augmented reality and virtual reality] collaboration in real time.”
He cited the example of moving from remotely assisted surgery in the near future to tele-operated surgery. “Our 5G labs are trying to predict the future, and we’re working with academia, startups, and companies to achieve it,” Ness said.
Verizon 5G challenges
To that end, Verizon Wireless has issued a number of challenges. In February, Verizon, Ericsson, and the Massachusetts Technology Leadership Council (MassTLC) announced the winners of the 5G Robotics Challenge at Verizon’s 5G Lab in Cambridge, Mass.
“We are at the threshold of a revolution,” said Blaze Vincent, senior customer solutions sales manager at Ericsson. “5G is part of a transformation for Industry 4.0, along with 3D printing, artificial intelligence and machine learning, and IIoT.”
Last month, Verizon hosted a 5G Demo Day, which included the following demonstrations by 5G Robotics Challenge winners:
- Ava Robotics Inc., whose mobile “Site Manager Robot” concept acted like a receptionist, directing people to meeting rooms and providing remote assistance via telepresence from a human receptionist using 5G. The company recently raised $2.9 million.
- RealBotics Inc., a Pittsburgh-based developer of software for remote control of robots and other machines using AR and VR, initially to enable training for complex machinery. 5G allows for remote video and operation in real time, said Chris Quick, RealTime Robotics CEO and a Robotics Summit speaker.
- Southie Autonomy, whose computer vision, software, and “wand” enable non-programmer users to instruct robots on what to do. “Usually, the cost of setting up a robot is three to five times that of the hardware,” said Rahul Chipalkatty, CEO of Southie Autonomy. “5G gives a 400% increase in speed.”
- The University of Massachusetts, Lowell, demonstrated a small humanoid robot whose motion planning was done on a MEC (multi-access edge computing or mobile edge computing) via 5G. The robot’s teleoperator sent high-level commands, such as direction or to climb some steps, but the actual joint commands were determined via the MEC, explained Dr. Yan Gu, an associate professor in UMass Lowell’s Department of Mechanical Engineering.
The following organizations also participated in the Demo Day at Verizon’s 5G Lab:
- Blueforce Development showed how its systems can gather biometric data, information about biological or chemical hazards, video, and location data from a variety of sensors and process them on the edge via 5G. This is based on military technology and could help police officers and other first responders with faster decision-making, said Bill Gellman, chief operating officer at Blueforce Development.
- Proximie demonstrated how surgeons could use AR over 5G to annotate high-resolution video from a remote operation to consult and improve quality of care. “There’s a global shortage of healthcare expertise,” said Wayne Kesti, North America business development manager at Proximie. “This allows surgeons to virtually scrub in from anywhere in the world.”
In addition to its 5G Robotics Challenge, Verizon ran an Ed Tech Challenge. At the Robotics Summit, Ness announced that the Built on 5G Challenge is open for submissions. Verizon is looking for small companies that are ready to scale, and it will help them with $1 million.
Qualcomm looks ahead to connected robots
“Industrial robots need 1 millisecond latency,” said Dev Singh, director of business development at Qualcomm Technologies during his closing keynote of Robotics Summit & Expo 2019. “More than 20 original equipment manufacturers have committed to launching 5G devices based on Qualcomm 5G technology.”
Qualcomm’s systems are getting smaller and will benefit from the ultra-reliable, low-latency connections. Developers can take advantage of these capabilities for applications as diverse as surgery and collaborative robots, said Singh.
“With wireless industrial Ethernet, we can have a reconfigurable factory,” he said. “There is strong industrial collaboration around 5G IoT, and Qualcomm is working with standards alliances.”
In response to questions, Singh said that Qualcomm is offering third-generation platforms for developers. He also said that some prospective users are more interested in removing cabling from industrial environments than in latency.
Factories and smart cities might need more smaller cell towers than a few large ones as for 4G, Singh acknowledged. “The cost of 5G towers will be balanced by the benefit of the returns on investment and six 9s of reliability,” he said.
Singh added that the power requirements between 5G and its predecessors are comparable. As for other device requirements, “we’re just beginning to peel the onion,” he said.
Another audience member asked how to get 5G into robots now. Singh said that modern processors are becoming available and that, as with autonomous cars, AI and cameras on the edge will put 5G and robots at the “tipping point,” he said, with research in one area helping others advance.
Shops with multiple devices that need to communicate and innovators looking for new robotics applications should get ready for 5G now, Singh said.
Filed Under: AI • machine learning, The Robot Report, Wireless, Robotics • robotic grippers • end effectors, Virtual reality