Rohde & Schwarz continues to push the boundaries of AI-driven wireless communication research with its latest milestone in neural receiver design and testing. To be showcased at MWC 2025 in Barcelona, the latest proof-of-concept, developed in collaboration with NVIDIA, integrates digital twin technology and high-fidelity ray tracing to create a robust framework for testing 5G-Advanced and 6G neural receivers under realistic propagation conditions. The objective is to bridge the gap between AI-driven wireless simulations and real-world deployment, thereby facilitating more efficient and accurate testing of next-generation receiver architectures.
This marks the fourth significant milestone in a long-term collaboration of Rohde & Schwarz with NVIDIA with demonstration showcased at past Mobile World Congress events. The joint demonstrations have pioneered advancements in AI/ML for wireless communications over the past years, from the training of neural receivers for handling analog impairments to the design of custom constellations for pilotless communications.
At the core of the demonstration is NVIDIA Sionna, a GPU-accelerated open-source library for link-level simulations, which provides ray-traced wireless channel models to generate realistic RF propagation conditions. The simulation results can then be seamlessly transferred to the R&S SMW200A vector signal generator from Rohde & Schwarz, which emulates complex real-world radio channels without requiring expensive external RF fading equipment. This advanced testbed enables testing and verification of AI/ML-based receiver algorithms and supports data-driven fine-tuning of neural components using realistic training data.
To ensure that the digital twin and ray-tracing models accurately reflect real-world conditions, the simulation is calibrated with data from a dedicated channel-sounding measurement campaign set in an urban street-canyon environment. By combining these precise measurements with NVIDIA Sionna, the ray tracer’s ability to model material interactions and electromagnetic propagation is further refined. The result is a calibrated version of the digital twin of the physical RF environment. This combination of technologies enables more accurate site-specific testing and validation of next-generation machine learning-based communication algorithms, including applications such as neural receivers and ML-based CSI feedback enhancements.
Gerald Tietscher, Vice President Signal Generators, Power Supplies and Meters at Rohde & Schwarz, said: “The collaboration with NVIDIA marks a significant milestone in advancing AI/ML applications in wireless communications. Our work has already demonstrated the potential of AI in wireless system design, from custom constellations to neural receivers handling real-world impairments. Now, with the integration of digital twins and ray tracing, we are further expanding the possibilities of AI-driven signal processing.”
Soma Velayutham, Vice President of Telecommunications at NVIDIA, said: “Digital twin technology has transformative potential in wireless system design. By integrating advanced NVIDIA ray tracing and machine learning into receiver development, Rohde & Schwarz is paving the way for AI-native 6G networks poised to offer outstanding efficiency and innovation compared with conventional implementations.”
Rohde & Schwarz is a key contributor to global 6G research, collaborating with industry partners, research institutions, and standardization bodies to develop the next generation of wireless technologies, expected to be commercially deployed by 2030
