A Lightweight Multi-Attack CAN Intrusion Detection System on Hybrid FPGAs

Shashwat Khandelwal,Shreejith Shanker

DOI: https://doi.org/10.1109/FPL57034.2022.00070

2024-01-19

Abstract:Rising connectivity in vehicles is enabling new capabilities like connected autonomous driving and advanced driver assistance systems (ADAS) for improving the safety and reliability of next-generation vehicles. This increased access to in-vehicle functions compromises critical capabilities that use legacy invehicle networks like Controller Area Network (CAN), which has no inherent security or authentication mechanism. Intrusion detection and mitigation approaches, particularly using machine learning models, have shown promising results in detecting multiple attack vectors in CAN through their ability to generalise to new vectors. However, most deployments require dedicated computing units like GPUs to perform line-rate detection, consuming much higher power. In this paper, we present a lightweight multi-attack quantised machine learning model that is deployed using Xilinx's Deep Learning Processing Unit IP on a Zynq Ultrascale+ (XCZU3EG) FPGA, which is trained and validated using the public CAN Intrusion Detection dataset. The quantised model detects denial of service and fuzzing attacks with an accuracy of above 99 % and a false positive rate of 0.07%, which are comparable to the state-of-the-art techniques in the literature. The Intrusion Detection System (IDS) execution consumes just 2.0 W with software tasks running on the ECU and achieves a 25 % reduction in per-message processing latency over the state-of-the-art implementations. This deployment allows the ECU function to coexist with the IDS with minimal changes to the tasks, making it ideal for real-time IDS in in-vehicle systems.

Cryptography and Security,Machine Learning,Systems and Control

What problem does this paper attempt to address?

The problem that this paper attempts to solve is the in - vehicle network security issues in modern vehicles due to increased connectivity, especially the security threats to the Controller Area Network (CAN). Specifically, the paper focuses on how to detect and defend against multiple attacks occurring on the CAN network, such as Denial - of - Service (DoS) attacks and Fuzzing attacks, through a lightweight multi - attack quantitative machine - learning model. These attacks take advantage of the weakness of the CAN network lacking inherent security or authentication mechanisms, which may lead to a complete interruption of vehicle functions or a complete loss of user privacy.

The paper proposes a lightweight multi - attack Intrusion Detection System (IDS) based on a hybrid Field - Programmable Gate Array (FPGA). This system is deployed on a Zynq Ultrascale+ (XCZU3EG) FPGA using Xilinx's Deep Learning Processing Unit (DPU) IP. By using the publicly available CAN intrusion detection data set for training and validation, this quantitative model can detect DoS and Fuzzing attacks with an accuracy of over 99% while maintaining an extremely low false - positive rate (0.07%). In addition, the execution of this system consumes only 2.0 watts of power, and compared to the state - of - the - art implementation methods, the processing delay per message is reduced by 25%. This deployment method allows ECU functions to coexist with the IDS with almost no modification to the tasks, which is very suitable for real - time intrusion detection in in - vehicle systems.