Project Abstract

The increasing sophistication of cyber threats to vehicle systems highlights the critical need for sophisticated security measures, notably within Controller Area Network (CAN) bus systems, which serve as the foundation of automotive communication infrastructures. This effort is motivated by the crucial need to protect these systems against both known and unknown abnormalities that may jeopardize vehicle safety and operation. This study’s distinctive value is its holistic approach to anomaly identification, which employs a combination of machine learning techniques such as Isolation Forest, Autoencoders, and Long Short-Term Memory (LSTM) networks. Each approach is chosen based on its individual capabilities in detecting various sorts of abnormalities, resulting in a robust solution adapted to the intricacies of CAN bus data.The research foundation for this study is a thorough assessment of genuine CAN bus traffic data to detect potential security flaws. This data is then utilized to do intensive feature engineering and train the suggested machine learning models, with the goal of accurately detecting abnormalities. The study’s design incorporates these several technologies to leverage on their unique advantages in anomaly identification.While the findings are expected, the end goal is to create a model capable of properly identifying a wide range of abnormalities in real time, considerably improving the security of automotive networks. This model tries to demonstrate a significant improvement in identifying small but essential abnormalities that earlier systems may have overlooked.The expected contribution of this research is a substantial breakthrough in car cybersecurity. If successful, this project would deliver a scalable and effective solution for real-time anomaly detection in CAN bus systems, filling a critical security gap in current automobiles. This will not only improve present understandings and approaches in vehicle network protection, but would also provide a new benchmark for future improvements in the sector.

Keywords: Artificial Intelligence, Data Science, Security

 Conference Details

Session: Presentation Stream 5 at Presentation Slot 1

Location: GH011 at Tuesday 7th 13:30 – 17:00

Markers: Nader Al Khatib (GTA), Tom Owen

Course: MSc Advanced Computer Science, Masters PG

Future Plans: I’m looking for work