EMMA-BEAM v0.2

Warning

Pre-implementation. This page describes proposed contracts. Class signatures, parameter types, schema fields, and behavior are subject to change before code lands. Once implementation exists, content here will be regenerated from docstrings or sourced from running tests.

Dataset. EMMA-BEAM-v0.2 covers the task, target domain, input, labels, splits, and the download command; this recipe pins only the rfgen configuration. Source. rfgen (primary).

rfgen recipe

  • rfgen commit: <rfgen-commit> (frozen at release; placeholder until pinned).

  • Emitters: the rfgen emitter_zoo group, comms family configured as 5G NR signals, placed across a deployment scenario.

  • Receiver array: massive-MIMO (multiple-input multiple-output) array, 32 or 64 elements, via the rfgen MultiRXConfig group. Element count and geometry pin at v0.2. Loaded into ArraySpec at load time.

  • Channel environment(s): E_URBAN (Sionna UMa), the dominant 5G deployment setting, pinned through the rfgen channel group.

  • Labeler + label extraction: the rfgen label group records per-beam received power; LabelExtractor maps the beam power vector to the best-beam index label.

  • SNR (signal-to-noise ratio) regime: drawn from the rfgen channel.snr_db_range.

  • Sample counts: <to be pinned at v0.2>.

  • Splits + seeds: train / dev / holdout seed ranges pinned at release; the holdout is a frozen secret seed, never published. OOD axis: SCENE_TYPE.

Real-capture counterpart

None at v0.2. Real massive-MIMO over-the-air beam captures are not available in the public testbed set; the sim-to-real gap for beam is asserted by recipe design.

Notes

  • Beam prediction depends on inter-antenna phase coherence, sharing the flagship’s load-bearing risk: rfgen’s array phase fidelity must be verified before the task is scored.

  • The OOD fold combines an unseen scenario with an unseen array geometry, which is the transfer the 3GPP AI/ML study targets but does not provide a dataset for.

References

  • 3GPP TR 38.843, “Study on Artificial Intelligence (AI)/Machine Learning (ML) for NR.” Ships beam-management use cases with no dataset or baseline. (verify revision-year)

  • NVIDIA Sionna: Hoydis et al., “Sionna,” 2023, arXiv:2203.11854. The UMa massive-MIMO channel model. (verify)

See Also