RSU-Based Online Intrusion Detection and Mitigation for VANET.

Area of Science:

• Cybersecurity
• Intelligent Transportation Systems (ITS)
• Machine Learning

Background:

• Secure vehicular communication is vital for intelligent transportation systems (ITS) and effective traffic management.
• Vehicular ad-hoc networks (VANETs) are vulnerable to integrity and availability attacks, including false data injection and stealthy distributed denial-of-service (DDoS) attacks.
• Current intrusion detection systems (IDS) require enhancement for timely and accurate threat identification in VANETs.

Purpose of the Study:

• To propose novel machine learning techniques for detecting and mitigating false data injection and stealthy DDoS attacks in VANETs.
• To leverage centralized communication via roadside units (RSUs) for enhanced security.
• To evaluate the performance of the proposed intrusion detection systems (IDS) against state-of-the-art solutions.

Main Methods:

• Development of novel machine learning algorithms for intrusion detection.
• Implementation of a centralized security architecture using roadside units (RSUs).
• Evaluation using a traffic simulator and a real-world traffic dataset.

Main Results:

• The proposed methods demonstrate superior detection and localization performance compared to existing solutions, with improvements ranging from 27% to 78%.
• The system maintains a comparable false alarm probability to state-of-the-art methods.
• Effective mitigation of false data injection and stealthy DDoS attacks was achieved.

Conclusions:

• The novel machine learning-based intrusion detection systems (IDS) offer significant improvements in VANET security.
• Centralized RSU-based security provides an effective framework for ITS.
• The proposed methods enhance the integrity and availability of vehicular communication networks.