Skip to main content

Wireless Benchmarks Datasets

This project tracks the datasets used to evaluate WavesFM and related baselines. Each dataset entry includes access links, citations, metrics, and TODOs to finalize splits and preprocessing so evaluations stay reproducible.

Available datasets (open from the sidebar):

  • RML2022 — automatic modulation classification.
  • POWDER RF Fingerprinting (RFP) — device-level identification with IQ captures.
  • ICARUS Interference (INTD / INTC) — interference detection/classification with IQ captures.
  • EfficientFi Human Activity Sensing (HAS) — WiFi CSI-based human activity sensing.
  • CommRad RF Signals (RFS) — spectrogram-based signal classification.
  • 5G NR Positioning (POS) — 5G NR CSI for high-precision positioning.
  • UWB Indoor Positioning and Tracking (UWB-Indoor) — UWB CIR-based localization and tracking.
  • UWB Industrial Localization (UWB-Industrial) — industrial UWB CIR-based localization.
  • RADCOM — modulation/signal type classification for radar and radio waveforms.
  • DeepMIMO (LoS/NLoS & Beam Prediction) — DeepMIMO channels for LWM downstream tasks.