5G will allow exporting robots’ AI into a nearby cloud

5G wireless connectivity is going to be very fast and wide enough to connect multiple devices in a reliable and flexible way. In fact, scientists say that 5G technology is what will allow inter-device communication, such as autonomous cars to communicate with one another and road infrastructure. Now a research team at New York university is building a 5G system, which will allow exporting robots’ artificial intelligence (AI) functions to the edge cloud.

5G technology offers more bandwidth and much higher speeds. Image credit: Fabian Horst via Wikimedia(CC BY-SA 4.0)

AI, obviously, offers a lot of advantages, particularly to robotics. It will allow machines to learn about their surroundings and adapt even if the environment is changing. However, one thing you rarely hear about AI technology is that it’s very heavy. These kinds of algorithms are highly demanding on hardware specification and, therefore, those computers can be pretty heavy. Not ideal for smaller robots. However, 5G may allow to export the AI capabilities to the edge cloud — the server in the cloud closest to the robot. This will make the robot lighter, but will still provide the benefits of AI without much latency thanks to the speed of 5G connectivity.

Of course, if that was easy, someone would have done it already. Scientists are currently creating algorithms that will allow this technology to take shape. Engineers have to solve issues of reliability, safety of robotic operation under communication degradation, and scalability to multi-robot systems. Exporting that capability to a cloud makes the system a bit more vulnerable for cyber attacks (who wouldn’t want to tap into advanced AI?) and communication break down. For example, what would happen to the robot if it lost its connection to the cloud where its intelligence is situated? Scientists have to seriously scratch their heads to make it a reliable and safe system.

For the longest time these ideas were more of a science fiction, because 4G didn’t have enough bandwidth and speed to support external AI systems. LIDAR, for example, requires quite a bit of bandwidth, but is also necessary for the robot to move around safely. Sundeep Rangan, one of the engineers in the project, said: “5G systems offer the potential for vastly higher data rates, but real-time cloud robotics remains challenging. A particular difficulty is that 5G communication using the millimeter wave (mmWave) bands are highly susceptible to blockage. As a result, links can have high peak data rates, but may be only intermittently available”.

The ultimate goal is lightweight, efficient AI-based robotics. This technology, obviously, would be useful in many different applications, but scientists are particularly interested in, well, science. It will create new opportunities to use AI cheaper and without wasting too much energy. 


Source: NYU

Leave a Reply

Your email address will not be published. Required fields are marked *