An Adversarial Attack Defending System for Securing In-Vehicle Networks

TL;DR

This paper investigates adversarial attacks on in-vehicle network security models, demonstrating high attack success rates and proposing a defense system that effectively detects such attacks with over 99% accuracy.

Contribution

The study introduces two new adversarial attack models and a novel defense system specifically designed for securing in-vehicle networks against these attacks.

Findings

Adversarial attacks succeed over 98% against LSTM-based detection.

Proposed AADS detects attacks with over 99% accuracy.

Focus on brake-related ECUs enhances security measures.

Abstract

In a modern vehicle, there are over seventy Electronics Control Units (ECUs). For an in-vehicle network, ECUs communicate with each other by following a standard communication protocol, such as Controller Area Network (CAN). However, an attacker can easily access the in-vehicle network to compromise ECUs through a WLAN or Bluetooth. Though there are various deep learning (DL) methods suggested for securing in-vehicle networks, recent studies on adversarial examples have shown that attackers can easily fool DL models. In this research, we further explore adversarial examples in an in-vehicle network. We first discover and implement two adversarial attack models that are harmful to a Long Short Term Memory (LSTM)-based detection model used in the in-vehicle network. Then, we propose an Adversarial Attack Defending System (AADS) for securing an in-vehicle network. Specifically, we focus on brake-related ECUs in an in-vehicle network. Our experimental results demonstrate that adversaries can easily attack the LSTM-based detection model with a success rate of over 98%, and the proposed AADS achieves over 99% accuracy for detecting adversarial attacks.