Applications¶
This section contains advanced, real-world examples and applications of DeepMIMO that go beyond basic tutorials. These examples demonstrate complete end-to-end workflows for specific use cases.
Available Applications¶
- Channel Prediction: Complete workflow for creating realistic channel sequences for ML-based channel prediction, covering interpolation techniques, Doppler effects, and comparative analysis.
- Sionna RT → DeepMIMO: Run Sionna RT 2.0 ray tracing on a built-in scene, export with
sionna_exporter, convert withdm.convert, and load the resulting dataset. Requiresdeepmimo[sionna]. - DeepMIMO → Sionna: Load a DeepMIMO scenario, adapt channels to Sionna's
(a, tau)format viaSionnaAdapter, build an OFDM channel matrix, and compute spectral efficiency. - OSM → Sionna RT → DeepMIMO: Generate a Mitsuba scene from OpenStreetMap building data — no Blender required — run Sionna RT, and convert to DeepMIMO format.
- Dynamic Ray Tracing: Run the ray tracer once per time snapshot to capture scenes where geometry itself changes (e.g. a moving transmitter or obstacle). Assembles snapshots into a
DynamicDataset. For most mobility scenarios, Tutorial 5 is faster and sufficient.
How Applications Differ from Tutorials¶
Tutorials focus on teaching specific DeepMIMO features and APIs in isolation.
Applications show complete end-to-end workflows for real use cases, often combining multiple features and demonstrating best practices for production scenarios.
Contributing¶
Have an interesting application or use case? We welcome contributions! See our contributing guide for details.