Source – https://www.eenewseurope.com/
Capgemini’s Project Marconi is using machine learning on a radio access network card using an Intel processor boosts spectral efficiency by 15 percent for low latency 5G applications
European consultancy Capgemini has developed a machine learning framework that works with OpenRAN hardware to boost spectral efficiency in 5G cellular networks
Project Marconi uses the Open Radio Access Network OpenRAN guidelines to maximize spectrum efficiency with real-time predictive analytics in a 5G Mediua Access Control (MAC) schedule. This is optimised for Intel’s AI Software and third generation Intel Xeon Scalable processors.
Capgemini used its NetAnticipate5G and RATIO OpenRAN platform to introduce advanced AI/ML techniques. The AI powered predictive analytical solution forecasts and assigns the appropriate MCS (modulation and coding scheme) values for signal transmission through forecasting of the user signal quality and mobility patterns accurately. The project improved AI accuracy to 55 percent and reduced AI inference time to 0.64msec, a 41 percent improvement.
In this way, the RAN can intelligently schedule MAC resources to achieve up to 40 percent more accurate MCS prediction which gives 15 percent better spectrum efficiency in testing. This is particularly important for applications that use low latency connectivity such as robotics-based manufacturing and V2X (vehicle-to-everything).
The Capgemini ML software on the Intel Xeon increases the amount of traffic each cell can handle and allows more subscribers alongside new Industry 4.0 services such as enhanced Mobile Broadband (eMBB) and Ultra Reliable Low Latency Communications (URLLC) use cases.
“Our teams worked closely with Intel to create a truly innovative solution that can really move the needle for operators,” said Walid Negm, Chief Research and Innovation Officer at Capgemini Engineering. “We gathered and utilized over one terabyte of data and conducted countless test runs with NetAnticipate5G to fine-tune the predictive analytics to meet diverse operator requirements. In short, machine learning can be deployed for intelligent decision-making on the RAN without any additional hardware requirement. This makes it cost efficient in the short run and future proof in the long run as we move into Cloud Native RAN implementations.”
“Our 3rd Gen Intel Xeon Scalable processors with built-in AI acceleration provide high performance for deep learning on the Net Anticipate 5G platform. Together, our collaboration delivered ultra-fast inference data to enhance the Open-Source ML libraries resulting in an intelligent RAN that can predict and quickly react to subscriber coverage requirements while reducing TCO,” said Cristina Rodriguez, VP of Wireless Access Network Division at Intel.
The €16bn Capgemini group has now integrated engineering group Altran that includes Cambridge Consultants, and has opened 5G R&D labs around Europe.