I. Introduction
In recent years, the automotive industry has seen rapid adoption of in-vehicle systems (such as infotainment system and advanced driver assistance system (ADAS)) and a multitude of sensors (such as a vision camera, radar, and lidar) that can accurately and quickly recognize the area around a vehicle. These advanced technologies generate a large amount of network traffic during communication between networks inside the system and communication with networks outside the vehicle. As a type of integrated sensing and communication (ISAC) concept, a new network environment is required to enable fast and accurate integrated processing of data of different types of protocols generated from heterogeneous networks of a vehicle [1]. Currently, many studies and industrial efforts toward handling a large amount of network traffic in an Automotive Ethernet are underway. Furthermore, to secure compatibility with existing web or mobile services, Automotive Ethernet adopts current ICT communication protocols, including the original Transmission Control Protocol (TCP)/Internet Protocol (IP). However, this indicates that various attack vectors and vulnerabilities that had occurred in the existing communication environment can be conveyed in the Automotive Ethernet environment. Even a minor risk posed by the attack vectors and vulnerabilities can jeopardize the safety of drivers, passengers, and pedestrians around vehicles at any time. Therefore, to ensure security of the IVN and the safety of drivers, a study needs to be conducted to detect the abnormality of a vehicle network and to identify their attack types in the Automotive Ethernet environment.