More Datasets

CARDINAL RF (CARDRF): An Outdoor UAV/UAS/DRONE RF Signals with Bluetooth and WiFi Signals Dataset

Lead Experimenter

Olusiji Medaiyese and Adrian Lauf, University of Louisville

Link to Dataset

IEEE Dataport

Equipment and Software Used

Keysight High-Sampling Oscilloscope, drones, remote controllers from different vendors, WiFi and Bluetooth interferers (this is a BYOD experiment)


The dataset contains UAV (telemetry and control), Bluetooth, and WiFi RF signals in .mat format. It was captured in an outdoor setting. Each RF signal has 5 million sampling points and spans a time period of 0.25ms. The script for plotting the signal is called SIGNAL_PLOT.mlx (a Matlab script) in the code zip file.


  1. O. O. Medaiyese, M. Ezuma, A. P. Lauf and A. A. Adeniran, "Hierarchical Learning Framework for UAV Detection and Identification," in IEEE Journal of Radio Frequency Identification, vol. 6, pp. 176-188, 2022, doi: 10.1109/JRFID.2022.3157653.
  2. O. O. Medaiyese, M. Ezuma, A. P. Lauf and I. Guvenc, "Semi-supervised Learning Framework for UAV Detection," in Proc. IEEE Annual International Symposium on Personal, Indoor and Mobile Radio Communications (PIMRC), pp. 1185-1190, Sep. 2021.
  3. O.O. Medaiyese, M. Ezuma, A.P. Lauf, and I. Guvenc, "Wavelet transform analytics for RF-based UAV detection and identification system using machine learning", Pervasive and Mobile Computing, 82, vol. 82, p.101569, June 2022.

Representative Results

The image below shows the data collection procedure and representative results on UAV/WiFi/Bluetooth classification accuracy. Additional results can be found in the references provided above.