A 24-hour signal recording dataset with labels for cybersecurity and IoT

Citation Author(s):
Yongxin
Liu
Embry-Riddle Aeronautical University
Jian
Wang
Embry-Riddle Aeronautical University
Houbing
Song
Embry-Riddle Aeronautical University
Shuteng
Niu
Embry-Riddle Aeronautical University
Thomas
Yang
Embry-Riddle Aeronautical University
Submitted by:
Yongxin Liu
Last updated:
Sat, 08/22/2020 - 13:03
DOI:
10.21227/gt9v-kz32
Data Format:
Links:
License:
0
0 ratings - Please login to submit your rating.

Abstract 

The dataset contains:
1. We conducted a A 24-hour recording of ADS-B signals at DAB on 1090 MHz with USRP B210 (8 MHz sample rate). In total, we got the signals from more than 130 aircraft.
2. An enhanced gr-adsb, in which each message's digital baseband (I/Q) signals and metadata (flight information) are recorded simultaneously. The output file path can be specified in the property panel of the ADS-B decoder submodule.
3. Our GnuRadio flow for signal reception.
4. Matlab code of the paper, wireless device identification using the zero-bias neural network.

Instructions: 

1. The "main.m" in Matlab code is the entry of simulation.
2. The "csv2mat" is a CPP program to convert raw records (adsb_records1.zip) of our gr-adsb into matlab manipulatable format. Matio library (https://github.com/tbeu/matio) is required.
3. The Gnuradio flowgraph is also provided with the enhanced version of gr-adsb, in which you are supposed to replace the original one (https://github.com/mhostetter/gr-adsb). And, you can specify an output file path in the property panel of the ADS-B decoder submodule.
4. Related publication: Zero-Bias Deep Learning for Accurate Identification of Internet of Things (IoT) Devices, IEEE IoTJ (accepted for publication on 21 August 2020), DOI: 10.1109/JIOT.2020.3018677