Andreas Ermedahl and Diarmuid Corcoran

At CASTOR Software days, they will talk about Five Realization Challenges of using AI for RAN Control – Experiences from Hands-On Prototype Research at Ericsson. 

A 5G Radio Access Network (RAN) is a highly complex distributed hardware and software combination. Optimal and efficient deployment and operation of 5G is a non-trivial task, and through the years considerable research- and development effort has been applied to this problem. AI- and data-driven machine learning (ML) techniques have shown considerable potential as a tool for developing and applying new resource management and control solutions for RAN. However, applying data-driven AI/ML techniques to a complex system like RAN brings unique realisation and implementation challenges. This talk lists some of these challenges and deep-dive into five.