Detecting Vehicle Anomaly in the Edge via Sensor Consistency and Frequency Characteristic

Guo, Fang; Wang, Zichang; Du, Suguo; Li, Huaxin; Zhu, Haojin; Pei, Qingqi; Cao, Zhenfu; Zhao, Jian-hong

doi:10.1109/tvt.2019.2907692

Cited by 47 publications

(21 citation statements)

References 22 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Also, the function is built on the Proportional Overlapping Scores approach that enables the amount of features contained in the Kyoto benchmark dataset to be reduced. Guo et al [28] suggested a new edge computing-based abnormal recognition strategy which employs edge-based sensor data fusion to identify anomalous events. The data from the sensor is used to find when the abnormality occurs inside the vehicle.…”

Section: Related Workmentioning

confidence: 99%

Sensor Data Based Anomaly Detection in Autonomous Vehicles using Modified Convolutional Neural Network

Rajendar¹,

Kaliappan²

2022

Intelligent Automation &Amp; Soft Computing

View full text Add to dashboard Cite

Section: Related Workmentioning

confidence: 99%

Sensor Data Based Anomaly Detection in Autonomous Vehicles using Modified Convolutional Neural Network

Rajendar¹,

Kaliappan²

2022

Intelligent Automation &Amp; Soft Computing

View full text Add to dashboard Cite

“…Moreover, the status features, i.e., the survival analysis of CAN message's frequency [7] and CAN bus's self-similarity [10], were utilized for the anomaly detection algorithms. In the application layer, in-vehicle sensors' correlation was used to check the abnormality with an On-Board Diagnostic (OBD) data [6]. Various deep learning algorithms based on increased computing power have recently been studied to detect invehicle network anomalies and intrusions [11].…”

Section: Intrusion Detection Techniquesmentioning

confidence: 99%

Car Hacking and Defense Competition on In-Vehicle Network

Kang¹,

Kwak²,

Lee³

et al. 2021

Proceedings Third International Workshop on Automotive and Autonomous Vehicle Security

View full text Add to dashboard Cite

in South Korea. The competition problem was to develop attacks and detection algorithms for CAN, a widely used standard of in-vehicle communication. The participants earned scores when they detect other team's attacks and inject critical or stealth attacks. The most significant difference from the events of Pwn2Own and SINCON is that we held attack and detection competitions simultaneously.The contributions of our competition are listed below:• We used a commercial car in the competition: Hyundai Avante CN7, a model released in 2020. The participants could inject attacks and see the following effects of a real vehicle. • The competition aimed to challenge participant's attacks and detection algorithms concurrently. On the day of the main contest, we captured the CAN traffic while the red team injects attack messages into the car and transmitted the traffic to the rest of the teams' (i.e., blue teams) detection systems. It is the first attempt to contest both attack and detection skills in the same car security contest to the best of our knowledge. • We ran a testbed with real vehicles before the finals, so even individual participants who do not have equipment could use the environment to inject and analyze CAN traffic. We believe this opportunity helped increasing interest in car security to researchers and students. • We could test and compare attack and detect methods of many research teams, including companies and universities in car security field. They showed statistical rulebased approach is effective to detect attacks, and reducing frequency of attack messages helps to avoid detection systems. II. BACKGROUND A. In-vehicle networkMost vehicles have standard protocols, i.e., LIN, CAN, MOST, and FlexRay, to communicate with in-vehicle nodes. In those protocols, the CAN protocol has high-integrity data communications for real-time applications, reliability, and excellent error detection. For such reasons, most common vehicles apply CAN to an in-vehicle network to transmit sensor data between the nodes: Electronic Control Unit (ECU), microcontrollers, and sensors [14]. It has characteristics such as multi-master bus access, bus topology, and message priority. An accessed CAN bus node can transmit CAN message without any configuration Abstract-Cybersecurity competitions can promote the importance of security and discover talented researchers. We hosted the Car Hacking: Attack & Defense Challenge from September 14, 2020 to November 27, 2020, and many security companies and researchers participated. To the best of our knowledge, it is the first competition to contest both attack and detection techniques on an in-vehicle network, specifically Controller Area Network (CAN). The participants developed various injection attacks and high-performance detection algorithms based on the real vehicle environment. Rule-based and ensemble tree-based models dominated the final round. Also, time interval and data byte patterns worked as major features to detect attacks.

show abstract

“…Anomaly detection in autonomous vehicles is not limited to the vision system (i.e., a extremely high data dimension problem). For example, non-vision low-dimension data from vehicle sensors are utilized in [25] to to detect anomalies from pair-wise data correlation, such as between the acceleration and wheel torque. The approach proposed by [26] is framed in the end-to-end autonomous driving context.…”

Section: 𝜕𝐼 𝜕𝑥 (𝑥) +mentioning

confidence: 99%

Improving Variational Autoencoder based Out-of-Distribution Detection for Embedded Real-time Applications

Feng

Easwaran

2021

ACM Trans. Embed. Comput. Syst.

View full text Add to dashboard Cite

Uncertainties in machine learning are a significant roadblock for its application in safety-critical cyber-physical systems (CPS). One source of uncertainty arises from distribution shifts in the input data between training and test scenarios. Detecting such distribution shifts in real-time is an emerging approach to address the challenge. The high dimensional input space in CPS applications involving imaging adds extra difficulty to the task. Generative learning models are widely adopted for the task, namely out-of-distribution (OoD) detection. To improve the state-of-the-art, we studied existing proposals from both machine learning and CPS fields. In the latter, safety monitoring in real-time for autonomous driving agents has been a focus. Exploiting the spatiotemporal correlation of motion in videos, we can robustly detect hazardous motion around autonomous driving agents. Inspired by the latest advances in the Variational Autoencoder (VAE) theory and practice, we tapped into the prior knowledge in data to further boost OoD detection’s robustness. Comparison studies over nuScenes and Synthia data sets show our methods significantly improve detection capabilities of OoD factors unique to driving scenarios, 42% better than state-of-the-art approaches. Our model also generalized near-perfectly, 97% better than the state-of-the-art across the real-world and simulation driving data sets experimented. Finally, we customized one proposed method into a twin-encoder model that can be deployed to resource limited embedded devices for real-time OoD detection. Its execution time was reduced over four times in low-precision 8-bit integer inference, while detection capability is comparable to its corresponding floating-point model.

show abstract

Detecting Vehicle Anomaly in the Edge via Sensor Consistency and Frequency Characteristic

Cited by 47 publications

References 22 publications

Sensor Data Based Anomaly Detection in Autonomous Vehicles using Modified Convolutional Neural Network

Sensor Data Based Anomaly Detection in Autonomous Vehicles using Modified Convolutional Neural Network

Car Hacking and Defense Competition on In-Vehicle Network

Improving Variational Autoencoder based Out-of-Distribution Detection for Embedded Real-time Applications

Contact Info

Product

Resources

About