Panel discusses how AI improves the benefits of Open RAN

Open RAN is entering the AI era.

The complex nature of tuning Open RAN radio heads for maximal coverage and minimal energy usage is an ideal application for AI. But there are a lot more applications now and the industry is taking notice.  

The industry’s standards bodies, 3GPP and the O-RAN ALLIANCE, are working together to develop standards for RAN optimization. But MNOs can experience a lot of the benefits of AI even while the standards efforts are in process.  

To explore these AI benefits, Javed Khan, our Senior Director, 5G Product Management joined an all-star industry panel at the recent Fierce Wireless Open RAN Summit 2024.  

Titled “AI-RAN: The Next Innovation in Radio Access Network Technology,” the session featured a keynote from Sebastián Guic, RAN Expert, Telecom Argentina.

On the panel, Khan was joined by Ravi P. Sinha, Co-Chairman, nGRG, O-RAN ALLIANCE, Lucia de Miguel, Senior Manager of Open RAN: Digital, Software & System Development, Vodafone Group, and Cristina Rodriguez, Vice President, Network & Edge Group & General Manager, Wireless Access Network Division, Intel. The panel was moderated by Stefan Pongratz, Vice President, Dell’Oro Group.  

What AI brings to Open RAN

The multiple benefits of AI in RAN were summarized by the keynoter as:

• Improved RAN performance as measured by improvement in metrics such as efficiency, capacity and radio performance.
• Improved quality of experience for consumers through the application of machine learning models that bring more network predictability.
• The agility that allows the optimization of resource allocation based on traffic behavior and the specific needs of applications on a per cell site basis.
• Reduced operational costs through the automation of network management and optimization tasks reducing the need for manual intervention.
• Flexible adaptation to new demands, innovations and emerging services to support in 5G, 5G Advanced and 6G.

Benefits increase over time

Khan added that the impact of these benefits will grow over time as AI models are trained on live data and can better anticipate changes in network data flows.

“Normally operators tend to dimension (the network) for the worst case or certain call models, but with AI/ML and the learning that gets introduced, you are able to more effectively predict network behavior, which you may not be able to do with existing algorithms,” Khan said.  

He cited the impact of load balancing as a use case.

“Load balancing, for example, is one use case where you're trying to predict where the UE is going to move and allocate resources or deallocate resources to serve the UE proactively instead of reactively. And so, the prediction is the key. And the learning that AI/ML is able to bring into the picture improves the ability to optimize it to a scale better than what traditional algorithms could potentially do.” 

There’s more from Javed Khan and all the panelists in the on-demand version of the webinar, which is available with a free registration from Fierce Wireless.

‍