Rohde & Schwarz is set to showcase a significant new milestone in AI-driven wireless system testing at MWC Barcelona 2026. Developed in a long-term collaboration with NVIDIA, the new testbed focuses on the prototyping and validation of AI-RAN innovations using cutting-edge test and measurement solutions.
Evolving from previous proof-of-concept designs—which focused on advanced neural receivers and custom constellations for pilotless communication—this new demonstration advances the technology from link-level validation directly to system-level verification using the full 5G NR protocol stack.
Powering Real-Time AI Inference in Wireless Systems
Powered by a single NVIDIA DGX Spark, the NVIDIA Sionna Research Kit runs a software-defined 5G RAN based on OpenAirInterface. Crucially, it supports AI inference workloads that comply with the strict, real-time constraints required by modern wireless systems.
To demonstrate the platform’s incredible flexibility, researchers have integrated a novel AI/ML-enhanced link adaptation algorithm into the end-to-end system. This algorithm dynamically adjusts the downlink modulation and coding scheme (MCS) to optimise spectral efficiency and link reliability.
This AI-driven link adaptation is capable of learning on the fly. It adapts not only to site-specific propagation characteristics but also to user equipment-specific behaviour, underscoring the critical need for end-to-end testbeds capable of capturing these complex, real-world effects.
Seamless Integration of Ray-Tracing and Channel Emulation
The cutting-edge testbed seamlessly integrates top-tier Rohde & Schwarz hardware with NVIDIA’s advanced software:
- R&S SMW200A Vector Signal Generator: Features dynamic channel emulation capabilities.
- R&S FSW Signal and Spectrum Analyser: Captures and analyses the complex RF signals.
- NVIDIA Sionna RT: A differentiable ray-tracing software that interfaces seamlessly with the physical instruments.
Together, this closed-loop setup enables researchers and developers to evaluate the performance of novel AI-driven RAN features under dynamic, site-specific RF conditions without ever having to leave the lab.
Bridging the Gap Between Simulation and Deployment
Industry leaders from both companies highlighted the importance of this milestone for the future of telecommunications.
“We’re excited to continue our ongoing collaboration with NVIDIA with this latest proof-of-concept for testing AI-enhanced base stations for both 5G-Advanced and 6G under realistic propagation conditions,” said Gerald Tietscher, Vice President Signal Generators, Power Supplies and Meters at Rohde & Schwarz. “Leveraging digital twin technology and ray tracing, this approach aims to bridge the gap between AI-driven wireless simulations and real-world deployment, facilitating more efficient and accurate testing of next-generation receiver architectures.”
Soma Velayutham, Global Industry Business Development Lead for Telecommunications at NVIDIA, added: “Synthetic data generation is transforming the way we train and validate AI-RAN systems by ensuring accuracy, scalability, and privacy, especially in settings of sparse data. Rohde & Schwarz, leveraging the NVIDIA Sionna Research Kit, exemplifies how industry-leading expertise and innovative technology can come together to accelerate progress in this critical field.”
Visitors to MWC Barcelona 2026 can experience the hardware-in-the-loop validation of these novel AI-RAN features live. Experts from both Rohde & Schwarz and NVIDIA will be available at the Rohde & Schwarz booth (5A80) in Hall 5 from 2 to 5 March 2026.
For more information on AI/ML for 6G networks, visit: www.rohde-schwarz.com/6G-AI-ML.
