Thierry E. KleinPresident, Bell Labs Solutions Research, Nokia Bell Labs
Whether it’s tasks humans can’t do or simply don’t want to do, robotic automation offers solutions — and 5G makes it feasible.
People often think of robots as singular machines, typically expensive and with limited applicability, such as the basketball-shooting robot at the Tokyo Olympics or the manufacturing robots on factory production lines. One-off devices like these put on impressive displays, but more is possible. Whether it’s anthropomorphic, a drone, a rover, or something else, 5G capabilities support revolutionary device-to-cloud and device-to-device communication that will make highly useful, intelligent robotics a reality.
Furthermore, by understanding the full capabilities of the next-generation networks and cloud technologies, device manufacturers can start building what’s possible when connecting the collective intelligence of devices in the cloud - unlocking a world of potentially infinite functionalities. At Nokia and through collaboration within the MITRE Engenuity Open Generation 5G Consortium, we are exploring the potential of machine orchestration for the benefit of humankind, and 5G serves as the powerful conductor of these modern orchestras of machines.
Re-Thinking Robots in a 5G Paradigm
Silo-designed robots geared toward specific tasks have inherent limitations, but if aligned in the cloud, robotics can provide businesses and society with a vast spectrum of benefits. Not only does moving a machine's intelligence to the cloud simplify individual devices and make them cheaper, it unlocks the collaborative abilities of singular robots, allowing them to work collectively and in sync to accomplish predefined and dynamic goals.
When considering the possibilities for robotics in the 5G era, we need to think about services, not just products. Picture this scenario: you're in New Jersey, and you want to visit Times Square. You type "Times Square" into your phone. Soon, an autonomous vehicle arrives at your location. You don't know how this car was chosen or how the car got here, but you get in, and the vehicle safely drives to Times Square and drops you off. You get billed automatically, and the payment automatically happens. It's super easy, and you don't have to worry about the details. You wanted to go to Times Square at 6 p.m. on Friday, and you did. Everything else is transparent to you, and it just works – you have been provided with a valuable robotic service. This same principle and simplicity and ease of use should be applied to a variety of robotic services in enterprise and industrial applications, smart cities, transportation and delivery use cases and consumer conveniences.
Thinking in these terms moves us beyond developing a specific robot or a device, like an autonomous car, and instead frames ambitions around functional deployment and orchestration of the devices in a broader context that benefits the public. To achieve this version of the future, we – the stakeholders, researchers, and architects of 5G connectivity – must think with vision, validate that 5G can support fleets of robots, and demonstrate that the devices are safe, viable, and beneficial.
Overcoming Obstacles Through Collaboration
Robot services simultaneously face two technical challenges: deploying an end-to-end 5G network and multi-vendor orchestration. Most device manufacturers haven’t opened up their interfaces to allow multi-robot, multi-vendor orchestration. As a result, startups and innovation arms of companies like Nokia resort to overlay software. Some will try to accomplish this entirely over the top, while others will establish business relationships with each robot manufacturer to access their proprietary information.
In the absence of openness from device manufacturers, it's primarily non-hardware companies working on this robot coordination. Ultimately, delivering these visionary end-to-end services requires a collection of players, because no single entity has access to all the pieces.
If your goal is delivering a service, that service requires the network, the user interface, and the payload application. Those elements are interdependent in terms of understanding how they will play off of each other. If the network operator doesn't know what you're going to use the robot for, it can't predict whether throughput, latency, or reliability is more important for you. It doesn't know what scale of deployment you need. Instead, it’s guessing or anticipating what your use cases are.
Without collaboration among all the players, the network can't be built in the most effective way. Vice versa, if you build a robot application and don't understand the capabilities, constraints, and limitations of the network, you may not be pushing your device to its potential because you might assume limitations that don't exist.
These interdependencies are why it's essential for network operators, device manufacturers, test range operators, and even end-user stakeholders to approach the acceleration of 5G together. Different parts of the ecosystem can build great things, but if we don't know all the capabilities and limitations of one another, we're not maximizing potential.
Use Cases that Drive the Most Benefit
Nokia is contributing to and actively working with members of MITRE Engenuity Open Generation to demonstrate the viability, safety, and benefits of orchestration over 5G. One focus involves uncrewed aerial systems (UAS), where we're identifying use cases that offer horizontal relevance, such as drone operations within indoor environments.
Whether it's a factory, a warehouse, or a stadium, in facilities where drones need to operate autonomously, you must be able to precisely locate, navigate and coordinate the drones. They need to be able to execute their mission without human intervention, they need to fly to a charging station when the battery level is low and they need to resume their operation after charging, all while coordinating their flights with other drones. And, of course, the drones must perform useful functions that serve a business purpose. Whether that function involves using a camera to identify an abnormality on a machinery, to detect a leak in a roof, to scan codes on packages in a warehouse or even to monitor plant growth in an indoor vertical farm, the system is effectively the same as far as the drone operation is concerned. The difference lies in the analytics being performed on the video and images streamed from the drone to the cloud.
By carefully selecting the use cases to test, we're aiming to solve many of the most complex yet foundational challenges to accelerating the 5G architecture that makes orchestration of robots and creation of robot services possible, whether those robots are autonomous taxis, pizza delivery vehicles, or inventory management drones. Across the ecosystem, the vision and technical capabilities necessary to realize these futuristic applications of devices are within reach. Through our work, we understand how to approach 5G more effectively while validating its revolutionary possibilities that will make our lives easier, more productive, safer, and more sustainable.