The head of InterDigital’s Wireless Lab weighs in on AI in a 6G context
Some hard questions are being asked of 6G these days: Is it inevitable? How risky will it be? Is 5G good enough for most of what most people want to do? And why would operators invest in 6G, when they are still working out how to effectively monetize 5G networks and get return on the billions that they’ve spent on 5G?
The standards work continues regardless, with studies for 6G standardization officially being kicked off at 3GPP’s meeting last month in Korea. And ETSI last week released a white paper detailing 18 use cases for one of the most-discussed 6G capabilities: Integrated and Sensing Communications (ISAC), which is sometimes called Joint Communications and Sensing (JCAS). That white paper offered up three potential integration levels and six possible sensing modes for ISAC, for consideration as part of 6G research and a framework for standardizatio.
In a similar vein, Milind Kulkarni, head of InterDigital’s Wireless Lab, told RCR Wireless News in an interview earlier this year at MWC Barcelona that 6G has been in a research and academic phase, but over the past year has begun to clarify in terms of a vision.
“Over the last year or so, it’s been now developing more and more towards, how do we realize that 6G vision?” Kulkarni said in an interview that was held just prior to the 3GPP meeting (in which InterDigital was a participant). Kulkarni expects that by the end of this year, 3GPP will have a consensus on the topics which will define 6G.
So this is the point at which the possibilities of 6G start to take a path to reality. As with every generation, some things will make it into the official standard; others will be discarded.
“As much as there is excitement about 6G, we can’t do everything,” Kulkarni acknowledged. “We have to be very selective.”
Artificial intelligence, he says, is a given. Beyond that, Kulkarni said, 6G must be about solving the issues that operators have had with 5G; establishing what 6G’s relationships and interoperability with 5G will be; and what new things it will bring that will excite consumers and industry.
Thinking more specifically about the likely frameworks for 6G that must be established, Kulkarni offered a number of aspects—most of which follow on 5G concepts. Those included:
-The continued evolution of MIMO.
-The continued push toward toward full-duplex operation, or using both the time domain and the frequency domain to gain more spectral efficiency.
-The continued evolution of terrestrial and non-terrestrial networks.
On top of those, he added AI—which is a given and has begun to emerge as a topic in the standards for 5G.
InterDigital’s half-dozen demos at this year’s MWC almost all included AI, from neural receiver model demonstration to better integrating AI into JCAS, and using AI to improve the handset experience while reducing energy use. Kulkarni also said that he had discussions around the use of ICAS for security—because RF can’t be blocked in the same way that a security camera can be—and potentially for things like combining sensing plus AI to, say, alert parents that they have accidentally left a child in a back seat in hot weather.
Even with that variety of potential applications, “it doesn’t even scratch the surface, to be honest, in terms of how AI can be used,” Kulkarni said. “AI will have multiple different uses, depending on whether you use it to increase efficiency, to optimize different things, to automate different things, or to solve some of the problems we haven’t been able to solve before. So we’ll see that in all the areas.” InterDigital is also focused on exploring AI-based services which leverage the network and the data that it collects.
Fundamentally, he said, the standards work has to first focus on how to create a platform for AI-enabled apps to operate on the network. “That that kind of comes down to all the data collection, data analytics, data management, and so forth. So once you have that, then you can build on it and say, okay, I can do automation of networks using that. I can learn the patterns of users, and I can put certain services to sleep so I conserve energy and then restart them based on the user patterns.
“Just about when you think you know the answer to something, somebody else comes up with a new innovation or a new app,” he said, adding: “So there are a few ideas that we have, there are a few ideas our peers have, or the industry has, but yet a lot to come.”
