Detecting In-vehicle CAN Message Attacks Using Heuristics and RNNs

In vehicle communications, due to simplicity and reliability, a Controller Area Network (CAN) bus is used as the de facto standard to provide serial communication between Electronic Control Units (ECUs). However, prior research reveals that several network-level attacks can be performed on the CAN bus due to the lack of underlying security mechanism. In this work, we develop an intrusion detection algorithm to detect DoS, fuzzy, and replay attacks injected in a real vehicle. Our approach uses heuristics as well as Recurrent Neural Networks (RNNs) to detect attacks. We test our algorithm with in-vehicle data samples collected from KIA Soul. Our preliminary results show the high accuracy in detecting different types of attacks.