Real-Time Intrusion Detection in Connected Autonomous Vehicles

Abstract

Connected autonomous vehicles (a.k.a CAVs) have shown a new paradigm in the transportation field by reducing cost, managing traffic and efficiently use of fuels. These developments have revolutionized not only the transportation field but also impacted our daily lives. However, growth in the field of automation has given rise to various security issues. Now, CAVs are using many sensors to perform automation. For proper navigation of vehicles, these sensor information must be communicated in a safe environment. In CAVs, these sensor information communicates through In-Vehicle-Network (IVN) to various parts of the vehicle, including different ECUs. But the network environment that has been used for communication is not always safe and can be infiltrated easily if proper security measures have not been taken. We are proposing a new intrusion detection system (IDS) to detect intrusions in real time in the network field of CAVs that is based on logical analysis of data (a.k.a LAD). For this, we have used the CAV-KDD dataset, which has been developed from a benchmark dataset KDDCUP’99 and resembles an actual network environment of CAVs. Our results have shown better performance than the existing results on CAV-KDD dataset.