MSc Thesis Proposal by Mahesh Abburi

Integrating Vehicular Ad-Hoc Networks (VANETs) into modern Intelligent Transportation Systems (ITS) has brought about significant advancements in transportation efficiency and safety. However, it has also introduced critical security concerns, particularly regarding the integrity of data exchanged among vehicles. This research focuses on tackling the emerging threat of Position Falsification Attacks in VANETs, where malicious entities broadcast fictitious location information to disrupt traffic flow and compromise road safety. Our methodology employs a detailed examination of the VeReMi dataset, a standard benchmark in VANET security research, alongside state-of-the-art machine learning classification algorithms. A key focus is not only on developing robust detection models but also on integrating XAI to enhance the interpretability of the outcomes. This approach ensures that the underlying decision-making processes of the ML models are transparent and understandable, fostering trust and facilitating more accessible validation by human experts.  Including XAI has demonstrated potential in providing deeper insights into model behaviours, particularly in understanding why specific predictions are made, thus identifying areas for model improvement. This thesis highlights the critical need to secure VANETs against Position Falsification Attacks and proposes an innovative solution by merging machine learning with explainable artificial intelligence. The findings contribute theoretically and practically, enhancing our understanding of VANET security challenges and providing actionable insights that can be implemented to safeguard vehicular communication networks against emerging cyber threats.

 

Keywords: VANETs, Machine learning, VeReMi Dataset, XAI, Position Falsification attacks, ITS.

 

Internal Reader:  Dr. Shaoquan Jiang

External Reader: Dr. Huapeng Wu

Advisor: Dr. Arunita Jaekel