The U.S. Department of Defense (DOD) is investing $2.7 million in a spectrum sharing project that will be administered by the Platforms for Advanced Wireless Research (PAWR). Although the project initially will focus on maximizing use of the CBRS band, it could have much broader applications, as a deputy director for the National Science Foundation (NSF) explained on a webcast today.
The PAWR is a project of Northeastern University and USIgnite that operates city-scale wireless testbeds. USIgnite is a nonprofit organization focused on smart communities with ties to the NSF.
As the government begins to run out of spectrum bands that can be put to commercial use, “we have to continuously invest in advanced research and development and look out for the frontier,” including spectrum sharing, said Thyagarajan Nandagopal, deputy director of NSF’s Division of Computing and Communications Foundations on today’s webcast, which was part of the 5GX Connect Virtual Summit.
Today, he said, “there are accepted models of spectrum sharing and there are many more that have not been explored.”
DOD Spectrum Sharing
The new project will focus on one of these unexplored, or only lightly explored, methods – the use of artificial intelligence (AI) to allocate spectrum resources.
As Nandagopal explained, this will be a departure from the way telecom network operators traditionally have managed networks.
“Telecom networks have been widely governed by manual processes,” explained Nandagopal. The typical 4G network “has around 5,000 parameters that somebody has to optimize,” he said.
The DOD spectrum sharing project aims to use AI to automate that process using technology developed by Zylinium Research.
“What Zylinium is trying to [find out] is ‘Can we use AI to make these decisions on the fly?’” said Nandagopal.
Spectrum sharing technology already plays a key role in the CBRS band by assigning commercial users to frequencies within the band that are not in use by the U.S. military. The PAWR project will take that concept further by enabling sharing within frequency bands. As a press release from the PAWR Project Office about the DOD spectrum sharing project explains, spectrum will be allocated at a highly granular level using resource blocks that measure 180 kHz by 1 ms.
In the future, Nandagopal noted that operators might use AI to share spectrum with one another. If one operator has a high traffic level, it might be able to use spectrum from another operator that has extra capacity that isn’t being used at that moment.
Considering today’s ultra-competitive wireless market, it may be hard to imagine network operators agreeing to such an arrangement with one another. But it’s important to recognize that “network operators” could include government users in frequency bands that are being eyed for potential commercial use. Perhaps those users could be persuaded to share the spectrum if AI proves to be an effective method for managing the sharing process.
As Nandagopal put it, the focus should be on “making sure the spectrum we have is used really well and making sure the networks we design are secure, reliable and robust.”