Spectrum mapping company Aurora Insight has a bird’s eye view of global 5G deployments, and from where CEO Jennifer Alvarez is sitting, it looks like the U.S. has a lot of ground to make up when it comes to mid-band deployments.
In January, the company launched a satellite named Charlie capable of creating a map of how spectrum from 1.4-7.1 GHz and 24-40 GHz is being used around the world. Though it’s only been a few months, Alvarez said Charlie is already yielding some fascinating insights.
“I would say that the pace of deployment is really what struck me, that in the matter of a few months you can see actual changes in how much spectrum is being used, whether it’s more bands coming online or just the intensities of the signals that we received in space, which are correlated with the number of transmitters, number base stations,” she said.
Alvarez noted data from Charlie showed mmWave 5G deployments across the globe are slow moving, while mid-band is being lit up at a near frantic pace. Rollouts of the latter are moving “very rapidly in Asia,” with a focus in China, South Korea and Japan. There, operators are “lighting up new deployments very, very frequently, hundreds of thousands of transmitters.” She added the U.S., which only recently began opening access to mid-band spectrum for 5G, is “way behind” on that front.
RELATED: U.S., China ‘race to 5G’ rages on
The company’s findings mesh with recent data coming out of China. ABI Research last month noted that as of October 2020, operators had deployed approximately 690,000 5G base stations in China. Last week, officials in the country said that figure rose to 792,000 by the end of February 2021.
Mid-band activity in the U.S. is set to heat up at the end of 2021 and into 2022 as recently auctioned C-band spectrum is made available. But Alvarez noted there is already some mid-band activity stateside in the 3.5 GHz band (also known as the Citizens Broadband Radio Service, or CBRS), which is shared between federal and commercial users. In particular, she noted there’s “a lot of activity in the middle” of the country, where those with CBRS licenses don’t have to worry about interference from federal entities.
Charlie has also provided a clearer picture of global connectivity gaps across different generations of cellular technology, with Alvarez noting the most prominent of these are in developing regions like Africa. As far as the continental U.S. is concerned on this front, Alvarez said there are gaps in both 4G and 5G coverage, but they don’t follow a set regional pattern.
“There’s not an area where you can say ‘this particular area, like the Midwest, is underserved.’ You can’t say that. It’s pockets,” she explained.
RELATED: CBRS users learn the ins and outs of Spectrum Access Systems
Though the number of base stations and thus coverage is largely correlated to population, Alvarez pointed out “you will see coverage in areas where there aren’t a lot of people but there’s a lot of industry. So, for example, in areas of west Texas where there’s fracking and other types of commercial activity, you’ll actually see quite a bit of wireless network activity even though there’s not a large population there.”
Bravo
Charlie is already generating plenty of data, but Alvarez noted Aurora Insight will soon have even more information to work with. The company is planning to launch a second satellite, Charlie’s twin Bravo, later this week (April 28).
With both Bravo and Charlie online, Alvarez said Aurora Insight will be able to produce 30 times more data than it was able to with its first satellite, Alpha, which launched in December 2018.
“They won’t be synchronous, they won’t be in the same orbit, they won’t be seeing the same things at the same time,” she said. “So it should be able to give us better information, a more complete picture of how the radio frequency spectrum is being used.”
Alvarez said data from its satellites is sent to the cloud alongside additional measurements gathered by plane and on the ground, where it is processed using machine learning. Given this, she said, more data is a good thing.
“Machine learning requires that the algorithm, the model, be trained on data. So a larger training set that’s more diverse will give you better, more generalized results that can be applied to different situations.”