In-Vehicle Network Intrusion Detection Using DNN with ReLU Activation Function

With the growing connectivity both within and between vehicles, there is an increasing focus on safety and security measures. Automotive systems rely on different serial protocols like Controller Area Network (CAN), Local Interconnect Network (LIN), and FlexRay. Among these, CAN bus is the predominant in-vehicle network protocol used for transmitting vehicle parameters among Electronic Control Units (ECUs). Unfortunately, the CAN protocol lacks inherent security mechanisms, rendering it susceptible to various types of attacks. Moreover, the expanded connectivity of vehicles has not only made the CAN bus vulnerable from within the vehicle but also from external sources.In this paper, we propose an intrusion detection system (IDS) based on Deep Neural Networks. The proposed system uses Rectified Linear Units (ReLU) as activation function, which avoid saturation and allow for better gradient flow in deep neural networks. The experimental investigation demonstrates that the proposed system gave superior results to existing algorithms.