RF Fingerprinting Migration
Authors/Creators
- 1. GOHM Electronics
Description
RF Fingerprinting Migration Dataset
1. Introduction
The RF Fingerprinting Migration Dataset is a collection of wireless signals designed to help researchers develop better ways to identify and authenticate Internet of Things (IoT) devices.
This work was supported by the European Union's Horizon Europe research and innovation programme under the ROBUST-6G project (Grant Agreement No. 101139068).
If you use this dataset in your research, please cite the following paper:
C. Ayyıldız, F. E. Yıldız, and B. E. Süzek, "Data-Efficient Domain Adaptation for Receiver-Invariant Radio Frequency Fingerprinting Identification," in 2026 Joint European Conference on Networks and Communications & 6G Summit (EuCNC/6G Summit), Málaga, Spain, 2026, doi: 10.1109/EuCNC/6GSummit68295.2026.11577304
2. Methodology
The dataset consists of 815,367 packets, totaling approximately 6.7 GB in size. Data collection took place in a laboratory environment, where the receivers were positioned next to each other, and the transmitters operated at distances of 0.5 m, 1 m, and 1.5 m.
Each transmitter transmitted 4,000 packets. Before storing the data, each captured packet was demodulated and decoded, and only successfully processed packets were retained. As a result, each packet has two identical copies received by the other receivers as complex64.
2.1 Receiver Specifications
The dataset was recorded using three Software Defined Radios (SDRs):
- R01: Fairwaves XTRX connected to Jetson AGX Xavier 01
- R02: Fairwaves XTRX connected to Jetson AGX Xavier 02
- R03: B200 Mini connected to NUC10i7FNH Workstation
- Identical antennas
- 400KHz sample rate
2.2 Transmitter Specifications
The transmission was done by 30 custom-built IoT transmitters.
- 2-GFSK Modulation
- 866MHz center frequency
- 50k baud rate
- 25KHz frequency deviation
- 100 ms duration within each packet
3. Packet Structure
Each packet has a fixed structure:
| Packet Feature | Data | Length (Bytes) |
| Preamble | 55 55 55 55 | 4 |
| Sync word | 93 0b 51 de | 4 |
| Length | 05 | 1 |
| Sequence Number | Various (0-4000) | 2 |
| Data | aa bb cc | 3 |
| CRC | Various | 2 |
4. Folder Structure
The dataset is organized in a hierarchical folder structure based on transmitters (Txx), distances (xx), and receivers (Rxx). This structure ensures that researchers can easily access data corresponding to specific transmitters, distances, and receivers.
The general folder structure is as follows:
├── Txx
│ ├── xx
│ │ ├── Rxx
Example Structure:
├── T01
│ ├── 05
│ │ ├── R01
│ │ ├── R02
│ │ └── R03
│ ├── 10
│ │ ├── R01
│ │ ├── R02
│ │ └── R03
...
5. Packet Naming Format
The dataset uses a simple naming system to label each packet. The format is:
Naming Format:Rxx_Txx_SequenceNumber_Distance.cfile
What Each Part Means:
-
Rxx(Receiver):- Shows which receiver collected the packet.
- Example:
R01means Receiver 1,R02means Receiver 2.
-
Txx(Transmitter):- Shows which transmitter sent the packet.
- Example:
T01means Transmitter 1,T20means Transmitter 20.
-
SequenceNumber:- The sequence number of the packet.
- Example:
300or1001.
-
Distance:- The distance between the transmitter and receiver.
- Example:
05is 0.5 meters,10is 1 meter, and15is 1.5 meters.
-
.cfile:- The file format used to save the data.
Examples:
R01_T01_300_05.cfile: Packet collected by Receiver 1 from Transmitter 1, sequence number 300, at 0.5 meters.R03_T20_1001_15.cfile: Packet collected by Receiver 3 from Transmitter 20, sequence number 1001, at 1.5 meters.
This naming system makes it easy to identify which receiver and transmitter were involved, as well as the packet’s sequence and the distance it was sent from.
6. Intended Use
This dataset aims to support the study of RF fingerprinting techniques. By using this data, researchers and developers can design more efficient methods, improving security against attacks. Ultimately, this leads to more sustainable and reliable IoT systems.
Files
rf_fingerprinting_migration.zip
Additional details
Identifiers
Related works
- Is cited by
- Conference paper: 10.1109/EuCNC/6GSummit68295.2026.11577304 (DOI)
Funding
Dates
- Available
-
2025-02-04Dataset available to download.