240 Validated RF Signal Samples Captured with RTL-SDR

240 validated RF signal samples were captured using an RTL-SDR Blog V4 dongle. The dataset is intended for wireless communication analysis, with a community notice indicating suboptimal SNR in some signal classes. Version 2.0 with improved captures is reported to be in development.

Use Cases

• Train a classifier on 240 RF signal samples to identify signal types, though it may learn frequency-specific noise characteristics as noted in the v1.0 notice.
• Analyze the signal-to-noise ratio (SNR) characteristics across different RF signal samples captured with the RTL-SDR hardware.
• Use the validated RF signal samples for benchmarking signal processing algorithms in wireless communication research.

Strengths

• Contains 240 validated RF signal samples, providing a concrete volume of data for analysis.
• Signals were captured using a specific, documented hardware device (RTL-SDR Blog V4 dongle).
• Dataset versioning and active improvement are indicated, with a v2.0 noted as in progress.

Limitations

• The dataset maintainer notes that several signal classes have suboptimal SNR, which may limit model generalization.
• With only 240 samples, the dataset is relatively small for training complex machine learning models without augmentation.
• The accuracy of the provided classifier is partially based on learning frequency-specific noise rather than pure signal features.

Provenance

Source

TrevTron on Hugging Face.

Collection Method

Captured using an RTL-SDR Blog V4 dongle.

Freshness

Last updated on 2026-03-19.