When it comes to things like self-driving cars, milliseconds matter. A UD researcher looks to speed up 5G
5G is fine for communication that isn’t trying to save your life. The latest iteration of wireless technology allows real-time communication for things like streaming videos, Facebook and Zoom. Anything under a 200-millisecond delay is considered adequate real-time communication.
University of Dayton researcher Feng Ye wants to get it down to 5 milliseconds by speeding up artificial intelligence (AI) algorithms. UD is partnering with the University of Nebraska-Lincoln and Utah State. They are recipients of a National Science Foundation grant.
Ye talked to WVXU Friday via Zoom. “I just tested between you and me right now; there is a communication latency of about 44 milliseconds," he says. "However, that is considered a super long latency in the 5G network. We really need it to be 5 milliseconds.”
Why does it need to be faster?
Ye points to something that must be precise, like remote surgery, for the increase in speed.
“So, let’s say I’m a doctor. I’ll give a command to the remote robot for instance, and then the robot will have about a 44 millisecond delay. If the needle travels at the same speed, it would have a 44 millisecond delay. That will be some operational error,” he says.
Also imagine a command center for autonomous driving.
“Let’s say we want to send a command to a car through LTE, there’s going to be a turnaround delay about 100 milliseconds, or 1/10th second. And if the car is traveling on the highway, that’s about 12 feet. So if there is some emergency, I’m not sure if 12 feet is going to be between life and death,” says Ye.
In smartphones and computers, AI algorithms work well, he says. But communication systems that include routers, switches and antennas are less capable devices.
His goal is to make 5G 100 times faster than existing Wi-Fi systems.