Verizon is using AI (artificial intelligence) models to help deploy thousands of 5G wireless transmitters for peak performance, The Wall Street Journal reports. The carrier will begin its multi-billion-dollar midband spectrum rollout later this year. The rollout will expand Verizon’s current ultra-wideband 5G wireless service. Verizon’s goal is to provide as much coverage with the least amount of transmitters. The small cell 5G transmitters can be put on vertical structures such as utility poles or buildings.
“When we build out these networks, these are very capital-intensive,” Verizon senior vice president and global chief information officer Shankar Arumugavelu told The Wall Street Journal. “We have to make sure that we are being very judicious in terms of how we are investing our capital.”
Verizon in-house data scientists and other employees are designing these models. They’re factoring in different variables that can alter the 5G signal’s strength such as buildings, bridges, weather, the transmitters’ location and other nearby transmitters. The carrier is trying to make up ground in its nationwide 5G deployment race with rivals AT&T and T-Mobile after a shortage of available airwaves for fast transmission and long signal ranges slowed it down. A lack of new network equipment being deployed also hindered Verizon’s 5G efforts.
Verizon’s first foray into using AI models happened in 2018. The carrier was trying to figure out the best spots to install transmitters for is ultra-wideband spectrum. It is turning to AI again for its midband spectrum, which is widely regarded as the wireless “sweet spot” for range and speed. Verizon anticipates it will cover 100 million people with midband spectrum by March 2022.
“Without AI, we would not be able to do this. Period,” Arumugavelu told The Wall Street Journal.
Verizon showed how serious it was about acquiring midband spectrum when it spent more than $50 billion to secure more than half of it in an FCC auction last year. The carrier plans to spend $10 billion in capital expenditures during the next three years to expand its midband range.
The Verizon AI system can find the best places to put the equipment by looking at a specific ZIP code and determining how well the signal could cover that area, Arumugavelu told The Wall Street Journal. The AI system’s computer-vision tech also lets the models analyze spatial and geographic elements, like bridges, trees or building heights in certain intersections, and locate the premiere spot for a 5G transmitter. The technology can also determine the best area for the antenna so it doesn’t interfere with other antennas in the area.
With billions of dollars at stake with the midband 5G spectrum, Verizon is relying on every asset it has at its disposal and that includes AI.
“(Verizon) has to be very precise and clinical to identify the places where they are going to place the nodes,” Gartner vice president of research for cloud and edge computing Sid Nag told The Wall Street Journal.
It’s imperative that Verizon, and other carriers, place 5G nodes in the perfect spot so their customers will be as happy as possible with their wireless service.
“Enterprises are hungry to use 5G,” Nag said.
Joe Dyton can be reached at joed@fifthgenmedia.com.
