Rohde & Schwarz and Qualcomm Validate Machine Learning-Based 5G Channel State Information Feedback

Munich – February 20, 2025 – Rohde & Schwarz and Qualcomm Technologies have announced the successful validation of machine learning (ML) based channel state information (CSI) feedback compression for 5G-Advanced networks. This breakthrough demonstrates a significant increase in throughput compared to conventional methods and confirms the feasibility of AI/ML implementation in wireless communications across different vendors.

The companies achieved interoperability between ML models running on Qualcomm Technologies' mobile device reference design and Rohde & Schwarz's CMX500 5G test platform. This enabled the implementation of enhanced CSI feedback mechanisms studied in 3GPP Releases 18 and 19. The successful validation resulted in a 51% improvement in throughput compared to previous standards.

The solution employed separate training approaches for the AI models on the network and device sides. Qualcomm Technologies developed a proprietary device encoder, while Rohde & Schwarz developed a decoder for its network emulator. The CMX500 test platform supports ML model integration through the ONNX format, allowing users to implement and validate their own AI architectures for wireless testing scenarios.

The significance of this validation lies in its demonstration that ML-based enhancements are effective and can be tested across equipment from different manufacturers. This represents a crucial step towards the commercialization of AI-driven wireless communication systems and lays the groundwork for future 6G networks, where AI is expected to be natively integrated into the air interface.

Rohde & Schwarz and Qualcomm Technologies will present this demonstration at Mobile World Congress 2025 in Barcelona.