2018 16th Annual Conference on Privacy, Security and Trust (PST) 2018
DOI: 10.1109/pst.2018.8514157
|View full text |Cite
|
Sign up to set email alerts
|

GIDS: GAN based Intrusion Detection System for In-Vehicle Network

Abstract: A Controller Area Network (CAN) bus in the vehicles is an efficient standard bus enabling communication between all Electronic Control Units (ECU). However, CAN bus is not enough to protect itself because of lack of security features. To detect suspicious network connections effectively, the intrusion detection system (IDS) is strongly required. Unlike the traditional IDS for Internet, there are small number of known attack signatures for vehicle networks. Also, IDS for vehicle requires high accuracy because a… Show more

Help me understand this report
View preprint versions

Search citation statements

Order By: Relevance

Paper Sections

Select...
2
1
1
1

Citation Types

1
262
0

Year Published

2019
2019
2024
2024

Publication Types

Select...
6
1
1

Relationship

0
8

Authors

Journals

citations
Cited by 355 publications
(263 citation statements)
references
References 9 publications
1
262
0
Order By: Relevance
“…To overcome these challenges, there should be more research on modelling CAN bus attacks like in [80] and creating attack databases like in [81,82]. Sharing datasets as an open-source (e.g., like in [66]) will help researchers; hence working on shared datasets will give a reference point to compare their research.…”
Section: Discussion On Can Security Researchmentioning
confidence: 99%
See 1 more Smart Citation
“…To overcome these challenges, there should be more research on modelling CAN bus attacks like in [80] and creating attack databases like in [81,82]. Sharing datasets as an open-source (e.g., like in [66]) will help researchers; hence working on shared datasets will give a reference point to compare their research.…”
Section: Discussion On Can Security Researchmentioning
confidence: 99%
“…Generative adversarial nets (GAN) based IDS [66] was proposed by Seo et al, who used the deep-learning model. The method is easy to expand and difficult to manipulate by an attacker, hence the detection mechanism has a black-box characteristic.…”
Section: Feature-based Idsmentioning
confidence: 99%
“…To evaluate the proposed IDS, this work considers two different datasets for the purpose of its implementation in both intra-vehicle and external networks as described in Section III-A. The first data set is called "Car-Hacking Dataset" or "CAN-intrusion dataset" which was proposed in 2018 for the purpose of IDS development on CAN bus [6]. On the other hand, to build a comprehensive IDS that can also be effective in external communication networks, a standard IDS data set containing the most updated attack scenarios, named "CICIDS2017" [5], is considered in this work.…”
Section: A Datasets Descriptionmentioning
confidence: 99%
“…One significant advantage of this approach is that attacks are effectively detected without changing the existing in-vehicle network structure. Seo et al [16] proposed a double discriminator using the generative adversarial networks (GAN). The method uses two different data sets to train two discriminators, which greatly improves the detection accuracy of IDS.…”
Section: Related Workmentioning
confidence: 99%
“…An extremely important problem is that the computational cost of machine-learning-based IDSs is considerably high. In contrast, we propose an IDS for in-vehicle network that can address these problems and the required computation shows advantage over preceding methods [16,17]. Table 2.…”
Section: Related Workmentioning
confidence: 99%