Creating digital representations of network and radio environments to evaluate configurations, algorithms and new technologies before deployment.
The Digital Twin platform simulates the whole stack: radio, propagation, orbits, and the data that trains the models. Site-specific propagation comes from 3D scenes through Sionna RT ray tracing, NTN scenarios cover LEO and GEO orbits, and multi-UE RF simulation runs many nr-UEs on one gNB over a simulated radio with no RF hardware needed.
Labelled per-UE, per-slot datasets are captured straight from the twin, so models can be trained on twin data, replayed on held-out scenarios, then deployed with a safe fallback.
Related
Related areas
AI RAN
Machine learning in the RAN scheduler
Per-UE intelligence, predictive link adaptation and dynamic scheduling inside the RAN scheduler.
Read more →Open NTN Testbed
Software testbed for satellite-enabled connectivity
Satellite-enabled connectivity explored end to end in software, with no satellite infrastructure required.
Read more →