A Study on Traffic Characteristics for Anomaly Detection of Ethernet-based IVN

In vehicles that provide intelligent services such as autonomous driving or ADAS, a wide bandwidth is required. CAN protocol, which is mainly used for a car network, has a bandwidth of only a few Mbps. So, there is an increasing trend to apply Ethernet to the in-vehicle network. The connected car with ICT technology increases the risk of intrusion by malicious viruses and external hackers. In this paper, we look at the stage of adopting Ethernet technology in vehicle network and examine the security threats of Ethernet-based vehicle networks. Also we look at the traffic characteristics of in-vehicle networks based on automotive ethernet. Finally, we briefly design an anomaly detection system using machine learning techniques.