Channel-state information (CSI) is crucial for maximizing user capacity and network performance in the rapidly changing 5G environment. Effective channel-dependent scheduling and adaptive modulation are made possible by CSI, guaranteeing reliable high-speed communications between the mobile device and base station. In 5G-Advanced and upcoming 6G networks, AI/ML-driven CSI improvements offer even higher efficiency, lower overhead, and better user experience.
It will be difficult to apply across vendors, though. By proving cross-vendor interoperability of ML-based CSI feedback upgrades, Rohde & Schwarz and Qualcomm Technologies have accomplished an industry milestone that will be displayed at MWC 2025 in Barcelona.
With assistance from Qualcomm Technologies, Rohde & Schwarz announced the successful validation of machine learning-based channel-state information feedback compression for 5G-Advanced networks, showcasing a notable boost in throughput over traditional techniques. This innovation demonstrates that implementing AI across vendors in wireless communications to improve network performance is feasible.
Through the use of improved CSI feedback methods included in 3GPP Releases 18 and 19, the two businesses were able to establish interoperability between ML models running on a mobile form factor reference design powered by a Qualcomm 5G Modem-RF and the Rohde & Schwarz CMX500 5G one-box signaling tester.
The configuration optimizes massive MIMO operations, which are essential for 5G networks, by enabling effective channel state compression based on CSI reference signal (CSI-RS) measurements.
According to 3GPP Release 15’s definition of Type I feedback followed by wideband precoding, this validation showed a 51% improvement in throughput performance.
Both parties used different methods for network-side and device-side AI model training. Through the use of certain reference models, compatibility was attained. Rohde & Schwarz created a decoder for its network emulator, and Qualcomm Technologies implemented a proprietary device encoder using autoencoder architecture.
Through the Open Neural Network eXchange (ONNX) protocol, the CMX500 one-box tester facilitates flexible ML model integration, allowing customers to deploy and verify their own AI architectures for wireless testing situations.
Cross-vendor interoperability as milestone for future standardization
Effective implementation and testing of two-sided ML-based air interface upgrades across equipment from several suppliers is demonstrated by the successful interoperability.
It offers a framework for testing and certification necessary for the commercial deployment of 5G-Advanced features, making it a major step toward adopting AI-enhanced wireless communications.
The impending standardization of a future 6G standard, where AI is anticipated to be natively embedded into the air interface architecture from the start, also marks this significant milestone. Therefore, future wireless systems will require cross-vendor AI interoperability as a fundamental component.
Leadership Comments
Christoph Pointner, Senior Vice President of Mobile Radio Testers at Rohde & Schwarz, says: “The capability of our CMX500 to implement and validate ML-based signal processing highlights the need for evolving test and measurement alongside wireless innovation. Thanks to ONNX support, customers can integrate their own ML architectures, making the CMX500 a versatile platform for AI-enhanced wireless testing. Validating interoperability with Qualcomm Technologies’ encoder confirms our commitment to establishing essential verification frameworks for AI-driven communications, from 5G-Advanced to future 6G systems.”
John Smee, Senior Vice President, Engineering, Qualcomm Technologies, Inc. adds, “In this new 5G-Advanced era of connectivity, and as we look toward 6G, AI is even more critical not just for the best user experiences but also for network performance. The joint research between Qualcomm Technologies and Rohde & Schwarz validates that AI-based CSI will help ensure these performance enhancements.”
Rohde & Schwarz will present the CSI feedback enhancements demonstration together with Qualcomm Technologies live at Mobile World Congress 2025 at Fira Gran Via in Barcelona in hall 5, booth 5A80.
A video preview of the demonstration is available here: https://www.rohde-schwarz.com/_251220-1545553.html
For further information on Rohde & Schwarz solutions for AI and ML in 6G networks: www.rohde-schwarz.com/6G-AI-ML
