Gated Recurrent Units for Intrusion Detection

Modern automobiles features a plethora of connected and smart features. This has lead to widespread growth of communications within a vehicle which has resulted in larger attack surfaces and different types of attacks. Classic cryptographic solutions are not suitable for the various sensors within a vehicle. Vehicle safety and timing to determine attacks is crucial. This paper proposes the use of Gated Recurrent Units (GRU) to perform intrusion detection in vehicles. Our proposed solution uses the temporal-based prediction properties of GRUs to determine if an intrusion is occurring in a vehicle. We present results of our experiments, conducted using real-world data, and analyze the findings.