DNVN - Keysight has just partnered with Samsung and NVIDIA to train artificial intelligence (AI) models for Samsung's 5G-Advanced and 6G technologies.
The performance of legacy RAN systems presents significant challenges for operators, including limited throughput, high latency, and reduced resource utilization. Commercial 5G deployments often require consideration of uplink performance, especially at the cell edge, where user equipment (UE) transmit power is limited, affecting the signal-to-noise ratio (SNR) received at the base station.
Due to high background noise, conventional channel estimation algorithms struggle in low SNR areas. The AI model provides a new approach by facilitating accurate channel estimation, optimizing resource allocation, and reducing power consumption. This can increase system capacity, improve network performance, and deliver a superior user experience.
This collaborative project enables Samsung to integrate a powerful AI model into its vRAN software solution.
Keysight’s channel simulation solutions provide channel generation capabilities for a wide range of channel conditions, along with real-time signal processing and radio frequency (RF) capabilities. Samsung has successfully created advanced AI models for channel estimation in uplink receivers, resulting in significant gains that can be observed in a lab environment. For example, simulation experiments have shown that throughput at the cell edge of the network is improved by up to 30 percent when using an AI model for channel estimation instead of the current static rule-based method.
A full-stack experimental setup, with radio and distribution equipment running on Samsung's NVIDIA Aerial AI platform, was used to evaluate the performance of this AI model. This AI model was deployed on the NVIDIA GH200 Grace Hopper Superchip platform, and Keysight's Channel Emulation and Core Emulation solutions.
This achievement will spur innovation and pave the way for widespread adoption of AI-enabled RAN technologies.
Van Thanh
Source: https://doanhnghiepvn.vn/cong-nghe/-huan-luyen-cac-mo-hinh-tri-tue-nhan-tao-cho-cong-nghe-5g-advanced-va-6g/20250319090520828
Comment (0)