Consumer Electronics Daily was a Warren News publication.
Beyond CBRS

AI-Based Spectrum Sharing Offers Alternative to Current Technologies

New, AI-driven technologies could offer an alternative to how spectrum sharing is done, experts said Wednesday during an RCR Wireless webinar. Panelists said AI could provide options to the citizens broadband radio service and increase dynamic sharing of government spectrum.

CTIA “has been very vocal” in stating concerns about the U.S. falling behind other countries on access to spectrum (see 2407250024), said Shaun McCarthy, president of Spectrum Effect, an AI-based wireless solutions company. “There’s a real concern that we’re falling behind our global peers.”

DOD uses the lower 3 GHz band, a top target of carriers. One solution would be for the military to vacate the band, McCarthy noted. But studies show that could cost up to $120 billion and take upward of 20 years, he said. “Clearly, we need a better solution.”

The solution is to use the radio access network as a “sensor” to detect incumbents' use, McCarthy said. “We think you need to use a wide-area, multi-site, intelligent analysis software platform that allows the RAN to dynamically maneuver around incumbent users.”

McCarthy said Spectrum Effect has been working on a solution with major carriers and Mitre and collected more than 7 million hours of data from operators around the globe to train AI models “to detect the very signatures that interference creates when it’s present in a mobile network” and then push configuration changes back to the network. Carriers are already reconfiguring their networks to address throughput and quality-of-experience issues, he said: What if reconfiguration was used “to maneuver around an incumbent?”

The presence of airborne radars in the lower 3 GHz band “considerably changes the dynamics” relative to CBRS, said Rob Soni, AT&T vice president-radio access network technology. As industry moves to more dynamic sharing, “we need to make sure that we have an efficient process ... for evaluating what technologies have been successful in the past” and what may not work.

Initial restrictions in CBRS probably limited the growth of that band, Soni said. The initial rules heavily restricted the output power, "which restricted the usage and made it less applicable as a macro-consumer wide-scale deployment frequency.”

The wireless industry was able to work quickly with the aviation industry to resolve C-band coexistence issues “and we see dramatic uptake in the usage” of that band, Soni said. AT&T and other carriers, working with DOD, have been able to use spectrum in the 3.45 GHz band, “and it works quite well.” The value of spectrum is closely correlated to the extent it can be used, he said.

One lesson learned from the administration’s America’s Mid-Band Initiative Team project (see 2010130033) was the importance of real-world testing before a band is opened for use by carriers, Soni said. “We need more spectrum, and there’s a desire to have more spectrum … without as many restrictions.”

Other countries are wrestling with the same sharing issues seen in the U.S., said Charles Clancy, Mitre chief technology officer, who was recently in the U.K. for meetings with regulators. “They have lots of questions about our experiences with CBRS because they’re trying to figure out how they’re going to manage their path to 6G.”

Harmonization of bands is of special interest in Europe because of the many shared borders, Clancy said. “The more harmonization we have internationally the better.”

DOD is interested in how it can use the same application programmable interfaces “to get access to spectrum no matter where they are in the world, particularly as we have a growing footprint in the Pacific and in Europe,” Clancy said. The more that systems are automated and DOD radars in the lower 3 GHz band can get the spectrum it needs without causing problems on the ground “is positive,” he said.

“We’re a long way from knowing exactly what the right answer is,” Clancy said: “We’re not ready to start writing a [3rd Generation Partnership Project] specification for how all of this should work. A lot more experimentation is needed.”