The Detection Method of Abnormal Messages based on Deep Neural Network on the CAN Bus

Chaoqun Xing

2019

Abstract

In recent years, with the increase of the number of external interfaces and electronic control units (ECUs) in vehicles, some unknown attacks about vehicle have emerged, the safety of vehicles has gradually become a top priority for auto manufacturer and vehicle owners. In this paper, we introduce a new attack model for the CAN bus of the vehicle and propose a deep neural network (DNN) based model to detect attacks. We collect normal messages exchanged between the CAN bus by ECUs in real vehicles to construct normal data sets. According to the existing methods, the ability of the attacker is quantified. Different anomaly data sets are constructed based on the strength of the attacker. Finally, the performance of the model is verified by using the normal and abnormal data sets. The detection rate of the proposed method is far more than the existing methods with good robustness and stability.

Download


Paper Citation


in Harvard Style

Xing C. (2019). The Detection Method of Abnormal Messages based on Deep Neural Network on the CAN Bus.In Proceedings of 5th International Conference on Vehicle, Mechanical and Electrical Engineering - Volume 1: ICVMEE, ISBN 978-989-758-412-1, pages 85-96. DOI: 10.5220/0008869100850096


in Bibtex Style

@conference{icvmee19,
author={Chaoqun Xing},
title={The Detection Method of Abnormal Messages based on Deep Neural Network on the CAN Bus},
booktitle={Proceedings of 5th International Conference on Vehicle, Mechanical and Electrical Engineering - Volume 1: ICVMEE,},
year={2019},
pages={85-96},
publisher={SciTePress},
organization={INSTICC},
doi={10.5220/0008869100850096},
isbn={978-989-758-412-1},
}


in EndNote Style

TY - CONF

JO - Proceedings of 5th International Conference on Vehicle, Mechanical and Electrical Engineering - Volume 1: ICVMEE,
TI - The Detection Method of Abnormal Messages based on Deep Neural Network on the CAN Bus
SN - 978-989-758-412-1
AU - Xing C.
PY - 2019
SP - 85
EP - 96
DO - 10.5220/0008869100850096