Requirements are specified in "requirements.txt"
$pip install -r requirements.txt
- Resonant sensor models are implanted in sensor driver.
- Attack parameters (induced frequency and attack amplitude) are connected to external commands.
- Mavlink module was modified to interface with the external commands.
- Additional logs : additional drone state variables, hardware timings are included.
- ROMFS (ROM Filesystem) file was revised to commnicate with inference computer.
- It provides software only testing.
$cd UnRocker_FW
$./UnRocker_Gyro_SITL.py or $./UnRocker_Accel_SITL.py
- Hardware related testing, so it requires the flight controller (FC).
$cd UnRocker_FW
$./UnRocker_Gyro_HITL.py or $./UnRocker_Accel_HITL.py
- Automated Dataset Generation is based on HITL mode.
$cd UnRocker_FW
$./UnRocker_TestsetGen_HITL.py
- Denoising AutoEncoder (DAE) is our core network.
$cd UnRocker_DAE_Gyro or $cd UnRokcer_DAE_Accel
$python3 train_conv_autoeoncoder.py
- Automatically generated dataset from HITL.
- Basic testset : Quadcopter Iris and Solo.
- Drone flight test data
- Actual injection test data
$cd UnRocker_DAE_Gyro or $cd UnRokcer_DAE_Accel
$python3 train_conv_autoeoncoder.py
- Realtime inference code based on TensorRT
$cd UnRocker_DAE_Gyro or $cd UnRokcer_DAE_Accel
$python3 trt_inference.py
@inproceedings{jeong2023unrocker,
title={UnRocking Drones : Foundattions of Acoustic Injection Attacks and Recovery Thereof},
author={Jeong, Jinseob and Kim, Dongkwan and Jang, Joonha and Noh, Juhwan and Song, Changhun and Kim, Yongdae},
booktitle={Network and Distributed Systems Security (NDSS) Symposium},
year={2023}
}
https://sites.google.com/view/unrocker/
e-mail: jeongjinseob@gmail.com