Ensemble and Gossip Learning-Based Framework for Intrusion Detection System in Vehicle-to-Everything Communication...

Area of Science:

• Cybersecurity in Intelligent Transportation Systems
• Machine Learning for Network Security

Background:

• Autonomous vehicles rely on vehicle-to-everything (V2X) communication for operation.
• V2X communication is vulnerable to cyber-attacks, particularly denial-of-service (DoS) attacks.
• DoS attacks can severely disrupt autonomous vehicle functionality and safety.

Purpose of the Study:

• To develop and evaluate methods for detecting DoS attacks in autonomous vehicles.
• To address DoS detection in both infrastructure-based (V2X Mode 3) and infrastructureless (V2X Mode 4) scenarios.
• To compare the proposed methods against existing security schemes.

Main Methods:

• Proposed an ensemble learning (EL) approach for DoS attack detection in V2X Mode 3 (infrastructure-based).
• Introduced a gossip learning (GL)-based approach for DoS attack detection in V2X Mode 4 (infrastructureless).
• Utilized the UNSW-NB15 dataset for evaluating the performance of both EL and GL methods.

Main Results:

• The ensemble learning approach achieved a DoS attack detection efficiency of 98.82%.
• The gossip learning approach demonstrated a DoS attack detection efficiency of 99.16%.
• Both proposed methods outperformed existing DoS detection schemes.

Conclusions:

• Ensemble learning and gossip learning are effective strategies for detecting DoS attacks in autonomous vehicle communication.
• The proposed methods offer robust security solutions for intelligent transportation systems.
• These findings contribute to the advancement of secure autonomous driving technologies.