The demonstration uses hardware-in-the-loop and compares a wireless receiver’s signal processing to traditional methods.
One of the many possible aspects of 6G swirls around AI/ML. Researchers are now exploring the possibility of using AL/ML to “teach” radios how to process signals. Rohde & Schwarz and NVIDIA are moving their work from simulations to implementing AI/ML for 6G. At MWC Barcelona, the companies will demonstrate a hardware-in-the-loop prototype of a neural receiver; one that shows achievable performance gains compared to traditional analog and digital signal processing.
MWC 2023 visitors can see a demonstration of how a neural receiver approach performs in a 5G NR uplink multi-user multiple input multiple output (MU-MIMO) scenario — a blueprint for a possible 6G physical layer. The setup combines high-end test solutions for signal generation and analysis from Rohde & Schwarz and the NVIDIA Sionna GPU-accelerated open-source library for link-level simulations.
A neural receiver replaces physical layer signal processing blocks with a wireless communications system that uses trained ML models. Researchers in research institutes, academia, and industry anticipate that a future 6G standard will use AI/ML for signal-processing tasks such as channel estimation, channel equalization, and de-mapping. So far, simulations suggest that a neural receiver will increase link quality and throughput compared to today’s high-performance deterministic software algorithms used in 5G NR.
Data sets are essential for training ML models. Often, the required access to data sets is limited or simply not available. Test and measurement equipment provides researchers with a viable alternative to that lack of data when generating data sets with different signal configurations to train machine learning models for signal processing tasks.
In the showcased AI/ML-based neural receiver setup at the Rohde & Schwarz booth, the R&S SMW200A vector signal generator emulates two individual users transmitting an 80 MHz wide signal in the uplink direction with a MIMO 2×2 signal configuration. Each user is independently faded, and noise is applied to simulate realistic radio channel conditions. The R&S MSR4 multi-purpose satellite receiver captures the 3 GHz transmitted signal through its four phase-coherent receive channels. The data then goes through a real-time streaming interface to a server. There, the digitized signal is pre-processed using the R&S Server-Based Testing (SBT) framework including R&S VSE vector signal explorer micro-services. The VSE signal analysis software synchronizes the signal and performs fast fourier transforms (FFT). This post-FFT data set serves as input for a neural receiver implemented using NVIDIA Sionna.
The demonstration compares the trained neural receiver to the classical concept of a linear minimum mean squared error (LMMSE) receiver architecture, which applies traditional signal processing techniques based on deterministically developed software algorithms.
Rohde & Schwarz will present the AI/ML based trained neural receiver demonstration at the Mobile World Congress 2023 at the Fira Gran Via in Barcelona.
Qiang Guo says
I am interested in seeing the video of this demo.