AI/ML can provide location data used to optimize wireless performance
Wireless networks are becoming increasingly complex and, with 5G looming, that will only continue as the mix and density of devices, as well as variability in application requirements, grows. A key to managing this complexity is the introduction of AI/ML tools that, over time, will slowly automate network operation and management.
At the recent Texas Wireless Summit, hosted by the University of Texas-Austin’s Wireless Networking and Communications Group, Intel’s Nageen Himayat discussed how AI/ML can help provide the context needed to dynamically optimize wireless systems.
“Like everybody in the industry and academia, we are also trying to comprehend how to best utilize machine learning for wireless applications.” She called out work being done within the WCNG to better understand using millimeter wave frequencies for vehicular communications, “which is a challenging problem because of high mobility and density of vehicular devices. You’re going to use contextual information such as positions of these vehicles to better do beam management–automatically predict which beam to use based on this location information.”
Here’s a video of Himayat’s talk.
Beam steering is a hot topic given the role millimeter wave will play in 5G deployments, particularly in support of early efforts to deliver enhanced mobile broadband. Essentially, because of its challenging propagation characteristics, a millimeter wave beam needs to target and track a handset or other piece of user equipment. The goal is to essentially make wherever the user is the center of the cell in terms of quality of experience.
Himayat explained that the use of AI/ML to connect a device with the “best cell or best beam” provides a number of benefits to both the user (lower latency) and the operator (lower power consumption).
Moving away from cars, she discussed work Intel has done around autonomous aerial vehicles.
“There are some challenges you have to address. When you’re high up in the sky…you see a lot more cells than you would normally see on the ground. That means you basically have more interference.”
The post Intel talks AI/ML for ‘context-aware wireless’ appeared first on RCR Wireless News.